Get the most recent info and news about Alphabet on HackerNoon, where 10k+ technologists publish stories for 4M+ monthly readers.

Alphabet Inc. provides online advertising services in the United States, Europe, the Middle East, Africa, the Asia-Pacific, Canada, and Latin America. It offers performance and brand advertising services. The company operates through Google and Other Bets segments. The Google segment offers products, such as Ads, Android, Chrome, Google Cloud, Google Maps, Google Play, Hardware, Search, and YouTube, as well as technical infrastructure. It also offers digital content, cloud services, hardware devices, and other miscellaneous products and services. The Other Bets segment includes businesses, including Access, Calico, CapitalG, GV, Verily, Waymo, and X, as well as Internet and television services. Alphabet Inc. was founded in 1998 and is headquartered in Mountain View, California.

Alphabet

alphabet

As it turns out, tech giant Alphabet (NASDAQ:GOOGL) (NASDAQ:GOOG) had plans for the money saved from layoffs, and those dollars are about to go to the UK. Once they...

Alphabet (NASDAQ:GOOGL) Prepares Hefty Investment in UK AI - TipRanks.com

While relatively attractively-priced when compared to other mega-cap stocks, some investors may remain wary of GOOG stock here.

Alphabet Outlook: My GOOG Stock Forecast for 2024 | InvestorPlace

Ensemble Capital Management, an investment management company, released its fourth quarter 2023 investor letter. A copy of the same can be downloaded here. The fund saw exceptionally excellent absolute and relative performance in the fourth quarter. The fund was up 16.00% in the quarter while ...

Here’s Why Investors Sold Off Alphabet (GOOG)

Google and Alphabet CEO Sundar Pichai has informed employees to brace themselves for further job layoffs throughout 2024.

Google mass layoffs: CEO tells employees to expect more job cuts in 2024

Google's parent company  Alphabet Inc., on Thursday. They were not an exception to the effect of the current market downturn affecting the world, as they saw revenue slum and growth slowed when compared to Refinitiv projections. For Q4, consolidated revenues were $76 billion, a 1% gain over Q4 in 2021. released its fourth and fiscal year 2022 results In the conference call, Alphabet CEO Sundar Pichai described the growth run during the pandemic as a period of significant acceleration in digital spending, which has made the current macroeconomic ecosystem challenging as the world returns to normality. As communities and business owners embrace artificial intelligence, the search and android giant is aligning itself to be a . The call was centered mainly on three things: the impact of AI, how to unlock its potential across its products, and how to focus investment to make necessary decisions as a company to get there. key player in the field “First, the AI opportunity ahead. AI is the most profound technology we are working on today. Our talented researchers, infrastructure, and technology make us extremely well-positioned as AI reaches an inflection point. More than six years ago, I first spoke about Google being an AI-first company.” Pichai said. According to , Google management had several meetings around its company AI strategy amidst the soaring popularity of OpenAI chatbot ChatGPT and the partnership of its search engine competitor Microsoft. During the call, Pichai talked about the company’s plans to release large language models (LLMs) like LaMDA and PaLM to the public in the coming weeks and months. a New York Times report LaMDA will be the first to go public as they will await feedback from people to improve. He described this model to be particularly amazing for composing, constructing and summarizing. The company will be incorporating these LLMs in companionship with search, as it will become more useful as they provide up-to-date factual information. The company also plans to introduce LLMs into two of its products — Gmail and Docs. Pichai said there are plans in motion to incorporate generative capabilities into its design segment. He further stated the company would provide tools and APIs for developers and creators to enable them to create their applications and explore the possibilities of AI on the platform. Google ads business is its biggest revenue source despite going down by 2% in the last quarter, coming in at $42.6 billion. According to senior vice president and Chief Business Officer Philipp Schindler, the company was able to outsize growth in advertising revenues during the pandemic, which saw advertising revenue of 2022 to be $90 billion higher than that in 2019. He reiterated the company's focus on generating revenue through AI-driven innovations, highlighting the progress made by smart bidding, which uses AI to predict future ad conversions and their value which help business to stay agile and responsive to a rapid shift in demand. The uses of AI advances boosted the bidding performance in the last year, which allowed advertiser outcomes down the funnel to drive ROI and use budgets more efficiently. “In 2022, AI advances boosted bidding performance, allowing us to move advertiser outcomes down the funnel to drive better ROI and use budgets more efficiently. In search query matching, large language models like MUM matched advertiser offices to user quarries. This understanding of human intent of language, combined with advances in bidding prediction, is why businesses can see an average of 35% more conversions when they upgrade exact match keywords to broad match in campaigns that use a target CPA. Google AI also underlies our creative products like tech suggestions in Google ads and creative optimization and responsive search ads.” Schindler said. Alphabet is throwing a strong signal to the AI industry after the company announced that they changing the financial report as it relates to its DeepMind AI in Q1. CFO Ruth Porat said this is due to increasing collaboration between DeepMind and Google products, and it will no longer be reported in other Bets and will now be reported as part of Alphabet’s corporate costs. Despite announcing a plan to and taking a severance charge of $1.9 to $2.9 billion, the company will be incurring an approximate sum of $500 million of costs related to exiting leases to align offices spaces with the adjusted headcount. Porat said the company would continue to hire in priority areas like artificial intelligence. She also said the company has a longer-term plan to re-engineer the cost base by incorporating AI to improve productivity across Alphabet for operational tasks. layoff 12,000 people in late January

Alphabet Doubles Up Efforts on AI as Tech Industry Races to Get a Foothold in the Sector

অ্যালফাবেট এআই-এর প্রচেষ্টাকে দ্বিগুণ করে কারণ টেক ইন্ডাস্ট্রি সেক্টরে পা রাখার জন্য রেস করছে

Alphabet verdoppelt seine Bemühungen im Bereich KI, während die Technologiebranche darum kämpft, in diesem Sektor Fuß zu fassen

Alphabet duplica los esfuerzos en IA a medida que la industria tecnológica se apresura a afianzarse en el sector

Alphabet redouble d'efforts sur l'IA alors que l'industrie technologique se précipite pour prendre pied dans le secteur

टेक उद्योग क्षेत्र में पैर जमाने के लिए अल्फाबेट ने एआई पर प्रयासों को दोगुना कर दिया

Alphabet が AI への取り組みを倍増

기술 산업이 해당 분야에서 발판을 마련하기 위해 경쟁함에 따라 Alphabet, AI에 대한 노력을 두 배로 늘림

Alphabet dobra esforços em IA enquanto a indústria de tecnologia corre para se firmar no setor

Teknoloji Sektörü Sektörde Yer Almak İçin Yarışırken Alphabet Yapay Zekaya Yönelik Çabaları İkiye Katlıyor

Bảng chữ cái tăng gấp đôi nỗ lực về AI khi ngành công nghiệp công nghệ chạy đua để có chỗ đứng trong lĩnh vực này

随着科技行业竞相在该领域站稳脚跟，Alphabet 加倍努力开发人工智能

Mark Twain's Speeches by Mark Twain, is part of the HackerNoon Books Series. You can jump to any chapter in this book . THE ALPHABET AND SIMPLIFIED SPELLING here THE ALPHABET AND SIMPLIFIED SPELLING ADDRESS AT THE DINNER GIVEN TO MR. CARNEGIE AT THE DEDICATION
          OF THE NEW YORK ENGINEERS’ CLUB, DECEMBER 9, 1907

          Mr. Clemens was introduced by the president of the club, who,
          quoting from the Mark Twain autobiography, recalled the day
          when the distinguished writer came to New York with $3 in small
          change in his pockets and a $10 bill sewed in his clothes. It seems to me that I was around here in the neighborhood of the Public Library about fifty or sixty years ago. I don’t deny the circumstance, although I don’t see how you got it out of my autobiography, which was not to be printed until I am dead, unless I’m dead now. I had that $3 in change, and I remember well the $10 which was sewed in my coat. I have prospered since. Now I have plenty of money and a disposition to squander it, but I can’t. One of those trust companies is taking care of it. Now, as this is probably the last time that I shall be out after nightfall this winter, I must say that I have come here with a mission, and I would make my errand of value. Many compliments have been paid to Mr. Carnegie to-night. I was expecting them. They are very gratifying to me. I have been a guest of honor myself, and I know what Mr. Carnegie is experiencing now. It is embarrassing to get compliments and compliments and only compliments, particularly when he knows as well as the rest of us that on the other side of him there are all sorts of things worthy of our condemnation. Just look at Mr. Carnegie’s face. It is fairly scintillating with fictitious innocence. You would think, looking at him, that he had never committed a crime in his life. But no—look at his pestiferious simplified spelling. You can’t any of you imagine what a crime that has been. Torquemada was nothing to Mr. Carnegie. That old fellow shed some blood in the Inquisition, but Mr. Carnegie has brought destruction to the entire race. I know he didn’t mean it to be a crime, but it was, just the same. He’s got us all so we can’t spell anything. The trouble with him is that he attacked orthography at the wrong end. He meant well, but he attacked the symptoms and not the cause of the disease. He ought to have gone to work on the alphabet. There’s not a vowel in it with a definite value, and not a consonant that you can hitch anything to. Look at the “h’s” distributed all around. There’s “gherkin.” What are you going to do with the “h” in that? What the devil’s the use of “h” in gherkin, I’d like to know. It’s one thing I admire the English for: they just don’t mind anything about them at all. But look at the “pneumatics” and the “pneumonias” and the rest of them. A real reform would settle them once and for all, and wind up by giving us an alphabet that we wouldn’t have to spell with at all, instead of this present silly alphabet, which I fancy was invented by a drunken thief. Why, there isn’t a man who doesn’t have to throw out about fifteen hundred words a day when he writes his letters because he can’t spell them! It’s like trying to do a St. Vitus’s dance with wooden legs. Now I’ll bet there isn’t a man here who can spell “pterodactyl,” not even the prisoner at the bar. I’d like to hear him try once—but not in public, for it’s too near Sunday, when all extravagant histrionic entertainments are barred. I’d like to hear him try in private, and when he got through trying to spell “pterodactyl” you wouldn’t know whether it was a fish or a beast or a bird, and whether it flew on its legs or walked with its wings. The chances are that he would give it tusks and make it lay eggs. Let’s get Mr. Carnegie to reform the alphabet, and we’ll pray for him—if he’ll take the risk. If we had adequate, competent vowels, with a system of accents, giving to each vowel its own soul and value, so every shade of that vowel would be shown in its accent, there is not a word in any tongue that we could not spell accurately. That would be competent, adequate, simplified spelling, in contrast to the clipping, the hair punching, the carbuncles, and the cancers which go by the name of simplified spelling. If I ask you what b-o-w spells you can’t tell me unless you know which b-o-w I mean, and it is the same with r-o-w, b-o-r-e, and the whole family of words which were born out of lawful wedlock and don’t know their own origin. Now, if we had an alphabet that was adequate and competent, instead of inadequate and incompetent, things would be different. Spelling reform has only made it bald-headed and unsightly. There is the whole tribe of them, “row” and “read” and “lead”—a whole family who don’t know who they are. I ask you to pronounce s-o-w, and you ask me what kind of a one. If we had a sane, determinate alphabet, instead of a hospital of comminuted eunuchs, you would know whether one referred to the act of a man casting the seed over the ploughed land or whether one wished to recall the lady hog and the future ham. It’s a rotten alphabet. I appoint Mr. Carnegie to get after it, and leave simplified spelling alone. Simplified spelling brought about sun-spots, the San Francisco earthquake, and the recent business depression, which we would never have had if spelling had been left all alone. Now, I hope I have soothed Mr. Carnegie and made him more comfortable than he would have been had he received only compliment after compliment, and I wish to say to him that simplified spelling is all right, but, like chastity, you can carry it too far. About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Mark Twain (2004). Mark Twain's Speeches. Urbana, Illinois: Project Gutenberg. Retrieved October 2022 https://www.gutenberg.org/cache/epub/3188/pg3188-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at , located at . www.gutenberg.org https://www.gutenberg.org/policy/license.html

THE ALPHABET AND SIMPLIFIED SPELLING

What Is Man? and Other Essays by Mark Twain, is part of the HackerNoon Books Series. You can jump to any chapter in this book . A SIMPLIFIED ALPHABET here A SIMPLIFIED ALPHABET (This article, written during the autumn of 1899, was about the last writing done by Mark Twain on any impersonal subject.) I have had a kindly feeling, a friendly feeling, a cousinly feeling toward Simplified Spelling, from the beginning of the movement three years ago, but nothing more inflamed than that. It seemed to me to merely propose to substitute one inadequacy for another; a sort of patching and plugging poor old dental relics with cement and gold and porcelain paste; what was really wanted was a new set of teeth. That is to say, a new . alphabet The heart of our trouble is with our foolish alphabet. It doesn’t know how to spell, and can’t be taught. In this it is like all other alphabets except one—the phonographic. That is the only competent alphabet in the world. It can spell and correctly pronounce any word in our language. That admirable alphabet, that brilliant alphabet, that inspired alphabet, can be learned in an hour or two. In a week the student can learn to write it with some little facility, and to read it with considerable ease. I know, for I saw it tried in a public school in Nevada forty-five years ago, and was so impressed by the incident that it has remained in my memory ever since. I wish we could adopt it in place of our present written (and printed) character. I mean the alphabet; simply the consonants and the vowels—I don’t mean any or abbreviations of them, such as the shorthand writer uses in order to get compression and speed. No, I would simply reductions spell every word out. I will insert the alphabet here as I find it in Burnz’s Phonic Shorthand. (Figure 1) It is arranged on the basis of Isaac Pitman’s . Isaac Pitman was the originator and father of scientific phonography. It is used throughout the globe. It was a memorable invention. He made it public seventy-three years ago. The firm of Isaac Pitman & Sons, New York, still exists, and they continue the master’s work. Phonography What should we gain? First of all, we could spell —and correctly—any word you please, just by the of it. We can’t do that with our present alphabet. For instance, take a simple, every-day word . If we tried to spell it by the sound of it, we should make it TYSIS, and be laughed at by every educated person. definitely sound phthisis Secondly, we should gain in in writing. reduction of labor Simplified Spelling makes valuable reductions in the case of several hundred words, but the new spelling must be . You can’t spell them by the sound; you must get them out of the book. learned But even if we knew the simplified form for every word in the language, the phonographic alphabet would still beat the Simplified Speller “hands down” in the important matter of economy of labor. I will illustrate: PRESENT FORM: through, laugh, highland. SIMPLIFIED FORM: thru, laff, hyland. PHONOGRAPHIC FORM: (Figure 2) To write the word “through,” the pen has to make twenty-one strokes. To write the word “thru,” the pen has to make twelve strokes—a good saving. To write that same word with the phonographic alphabet, the pen has to make only strokes. three To write the word “laugh,” the pen has to make strokes. fourteen To write “laff,” the pen has to make the of strokes—no labor is saved to the penman. same number To write the same word with the phonographic alphabet, the pen has to make only strokes. three To write the word “highland,” the pen has to make twenty-two strokes. To write “hyland,” the pen has to make eighteen strokes. To write that word with the phonographic alphabet, the pen has to make only FIVE strokes. (Figure 3) To write the words “phonographic alphabet,” the pen has to make fifty-three strokes. To write “fonografic alfabet,” the pen has to make fifty strokes. To the penman, the saving in labor is insignificant. To write that word (with vowels) with the phonographic alphabet, the pen has to make only strokes. seventeen Without the vowels, only strokes. (Figure 4) The vowels are hardly necessary, this time. thirteen We make five pen-strokes in writing an m. Thus: (Figure 5) a stroke down; a stroke up; a second stroke down; a second stroke up; a final stroke down. Total, five. The phonographic alphabet accomplishes the m with a single stroke—a curve, like a parenthesis that has come home drunk and has fallen face down right at the front door where everybody that goes along will see him and say, Alas! When our written m is not the end of a word, but is otherwise located, it has to be connected with the next letter, and that requires another pen-stroke, making six in all, before you get rid of that m. But never mind about the connecting strokes—let them go. Without counting them, the twenty-six letters of our alphabet consumed about eighty pen-strokes for their construction—about three pen-strokes per letter. It is required by the phonographic alphabet. It requires but stroke for each letter. three times the number one My writing-gait is—well, I don’t know what it is, but I will time myself and see. Result: it is twenty-four words per minute. I don’t mean composing; I mean . There isn’t any definite composing-gait. copying Very well, my copying-gait is 1,440 words per hour—say 1,500. If I could use the phonographic character with facility I could do the 1,500 in twenty minutes. I could do nine hours’ copying in three hours; I could do three years’ copying in one year. Also, if I had a typewriting machine with the phonographic alphabet on it—oh, the miracles I could do! I am not pretending to write that character well. I have never had a lesson, and I am copying the letters from the book. But I can accomplish my desire, at any rate, which is, to make the reader get a good and clear idea of the advantage it would be to us if we could discard our present alphabet and put this better one in its place—using it in books, newspapers, with the typewriter, and with the pen. (Figure 6)— . I think it is graceful and would look comely in print. And consider—once more, I beg—what a labor-saver it is! Ten pen-strokes with the one system to convey those three words above, and thirty-three by the other! (Figure 7) I mean, in SOME ways, not in all. I suppose I might go so far as to say in most ways, and be within the facts, but never mind; let it go at . One of the ways in which it exercises this birthright is—as I think—continuing to use our laughable alphabet these seventy-three years while there was a rational one at hand, to be had for the taking. Man Dog Horse some It has taken five hundred years to simplify some of Chaucer’s rotten spelling—if I may be allowed to use so frank a term as that—and it will take five hundred more to get our exasperating new Simplified Corruptions accepted and running smoothly. And we sha’n’t be any better off then than we are now; for in that day we shall still have the privilege the Simplifiers are exercising now: can change the spelling that wants to. anybody It will always follow the SOUND. If you want to change the spelling, you have to change the sound first. But you can’t change the phonographic spelling; there isn’t any way. Mind, I myself am a Simplified Speller; I belong to that unhappy guild that is patiently and hopefully trying to reform our drunken old alphabet by reducing his whiskey. Well, it will improve him. When they get through and have reformed him all they can by their system he will be only HALF drunk. Above that condition their system can never lift him. There is no competent, and lasting, and real reform for him but to take away his whiskey entirely, and fill up his jug with Pitman’s wholesome and undiseased alphabet. One great drawback to Simplified Spelling is, that in print a simplified word looks so like the very nation! and when you bunch a whole squadron of the Simplified together the spectacle is very nearly unendurable. The da ma ov koars kum when the publik ma be expektd to get rekonsyled to the bezair asspekt of the Simplified Kombynashuns, but—if I may be allowed the expression—is it worth the wasted time? (Figure 8) To see our letters put together in ways to which we are not accustomed offends the eye, and also takes the out of the words. expression La on, Makduf, and damd be he hoo furst krys hold, enuf! It doesn’t thrill you as it used to do. The simplifications have sucked the thrill all out of it. But a written character with which we are does not offend us—Greek, Hebrew, Russian, Arabic, and the others—they have an interesting look, and we see beauty in them, too. And this is true of hieroglyphics, as well. There is something pleasant and engaging about the mathematical signs when we do not understand them. The mystery hidden in these things has a fascination for us: we can’t come across a printed page of shorthand without being impressed by it and wishing we could read it. not acquainted Very well, what I am offering for acceptance and adoption is not shorthand, but longhand, written with the . You can write three times as many words in a minute with it as you can write with our alphabet. And so, in a way, it properly a shorthand. It has a pleasant look, too; a beguiling look, an inviting look. I will write something in it, in my rude and untaught way: (Figure 9) Shorthand Alphabet Unreduced is Even when do it it comes out prettier than it does in Simplified Spelling. Yes, and in the Simplified it costs one hundred and twenty-three pen-strokes to write it, whereas in the phonographic it costs only twenty-nine. I (Figure 9) is probably (Figure 10). Let us hope so, anyway. About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Mark Twain (2004). What Is Man? and Other Essays. Urbana, Illinois: Project Gutenberg. Retrieved October 2022 https://www.gutenberg.org/cache/epub/70/pg70-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at , located at . www.gutenberg.org https://www.gutenberg.org/policy/license.html

A SIMPLIFIED ALPHABET

If you haven't realized it yet, this is a crazy market! However, beyond short-term volatility driven by liquidity and macroeconomic trends. There are a few interesting takes from the latest financial releases of Google, Facebook, and Amazon. Let's see some of them! Google’s parent company, Alphabet, released its end-of-year financials. Looking at its financials is important for various reasons and some of them are: Assessing the digital advertising landscape. Understanding how Google’s advertising machine works at scale. How consumers’ behavior is shifting in this new digital landscape. Let’s dive into them. Google’s profitability Google’s profitability has slightly improved in 2021, thanks to the fact it managed to increase its faster than it increased its operational costs, as traffic on its platforms increased substantially during 2020-2021. We’ll see in this in-depth why that’s the case. revenue analysis First of all, for the first time in its history, Google generated over $257 billion in revenues. And the company almost reached a two trillion market capitalization. For some context, when Google IPOed in 2004, it recorded almost a billion in revenues, and it was worth about $23 billion, as it popped at its IPO date to $85 per share (On February 2nd, 2022, a Google stock is worth $2,958.22). At the time, in 2004, Google had just managed to scale its advertising machine comprised mainly of Google AdWords (today Google Ads) and Google AdSense. At the time the advertising machine was primarily based on Internet traffic from desktop devices. That was a completely different world. As we’ll see throughout this , today most traffic comes from mobile. And Google ’s mobile ads platforms (Google AdMob) play a key role. So let’s dive a bit into the main segments of . analysis financial Google Google’s main segments in 2021 In 2021, the Google advertising machine, generated over $209 billion in revenues. This represented an over 42% , year over year! This is a massive improvement for a company worth almost two trillion. In a market landscape that is not anymore in favor of advertising. How did Google pull this off? growth digital We’ll see this shortly. But for now, let’s emphasize a few key points. Today the Google advertising machine is comprised of three main products: : This represents the set of products that Google owns, from the search engines to all the other vertical platforms that the company operates (Google Discover, Google News, Google Travel, and more). In 2021, this segment generated over $148 billion! Google Search/Properties : This is of course one of the most successful acquisitions ever done. It was acquired by Google for $1.65 billion in 2006. It’s important to notice that Google was able to integrate YouTube and scale it up. A feat that not every other company would have been able to achieve. At the time YouTube was getting sued for various copyright infringements (the is comprised of user-generated content, often posting copyrighted materials) that would have most probably bankrupted it if it had stayed a startup without Google’s backing. By 2021, YouTube has become an advertising machine generating over $28 billion (this doesn’t count the YouTube memberships, which are reported separately). YouTube business platform this is the set of publishers that decide to opt into Google’s advertising network (either AdSense for desktop, or AdMob for in-app advertising). Here Google shows advertising on the network members’ properties, thus splitting the revenues with them. In 2021, Google’s network members generated over $31 billion of revenues. And Google Network Members’ properties: As we’ll see understanding the difference among these segments helps us understand how Google manages its cost structure for each segment. What determines the rowth of each segment? g In fact, since the pandemic hit, more and more users started to use Google’s products. This trend has continued. However, most of it was driven by mobile users’ rowth . This is an important aspect, as it shows that Google’s main driver of rowth is based on mobile traffic. This changes the way the company to prioritize its developments efforts, its ad formats served to users, and also how it experiments. Google’s search advertising has been driven by g rowth in search queries. g g needs product This means that Alphabet (as it’s evident to anyone going on YouTube) has ramped up the advertising operations on YouTube. In short, on YouTube now there are way more ads than before. This “improved ad formats” is the result of YouTube’s extreme stickiness with users, which enables Alphabet to play with its ad formats. YouTube g rowth was driven primarily by improved ad formats. In short, the mobile advertising platform, powered up by Android devices through the Google Play store, was the main driver of in 2021. This shows how Google has shifted also its focus on the mobile advertising platform. Google’s network members’ properties were primarily driven by AdMob. revenue growth Let’s see how, a little bit more in detail, ad monetization changed for Google. In 2021, two main factors determined the improvement in ad monetization by Google. First, Google recorded an increase in paid clicks (driven by an increase in user adoption and search queries primarily on mobile devices). Second, this also drove more paid clicks in AdMob through the Google Play store. Third, as we saw Google is testing various ad formats (we can argue it’s showing more ads) both on Google’s products and YouTube, which slightly improved monetization. Among the other revenues, instead, Google Cloud also recorded an important thanks to the Google Cloud Platform. growth Let’s look now at the other side, the cost structure. Google’s cost structure When it comes to Google its cost structure is pretty straightforward. In order to keep its operations profitable, it needs to be able to generate traffic at much lower costs, than it can monetize it. In 2021, Google traffic acquisition costs (the spending needed to run the servers, for its main products, and the deals in place to enable the for its main products) were over $45 billion. distribution In respect to advertising revenues, its traffic acquisition costs were 22% of its advertising revenues in 2020. In 2021, instead, the traffic acquisition costs decreased to 21.7% compared to its revenues. This might seem a little difference. Yet, Google managed to lower its traffic acquisition costs, nonetheless its much wider users’ adoption. This shows that a lot of that adoption was organic, and based on strong deals the company has in place. The other cost of revenues instead was mainly driven by improved content acquisition costs and costs in data centers. In terms of profitability, it’s interesting to notice, how, at a wider scale, Google Services has huge margins. That’s because as we saw in the previous paragraph, Google managed to improve monetization for users, it saw mobile users’ growth, while it managed to lower its traffic acquisition costs! This combination led to an improved marginality. However, it’s also interesting to notice how the Google Advertising machine is the only one running at positive margins. Where instead, both the Google Cloud platform and the other Google Bets run at negative margins. Important to distinguish here. The Google Cloud platform is critical for the future success of the Google AI platform. When it comes to the other bets those instead are breakthroughs that Alphabet is pursuing with a long-term perspective. Those are money-losing bets for now but might turn into something interesting in the coming decade. Back in the early 2000s, the company had started to build its advertising machine. This journey culminated in 2004 when the company had managed to put together the various pieces (Google AdWords and Google AdSense) of its advertising machine, thus scaling up its revenues on top of a growing search platform. Indeed, in 2004 Google generated almost a billion dollars in revenues, and it was worth about $23 billion as it IPOed. In 2021, the company passed the $257 billion mark! Google’s Alphabet has become a two trillion-dollar company in 2021. This trend continues across 2020-2021. The increased search query traffic was primarily due to mobile users. Therefore, Google had to push more experimentations through its mobile platforms. Among the biggest drivers of its revenues were the increased adoptions from users as the pandemic hit. Alphabet here pushed through more ad formats, and more advertising to monetize it further. Thus, taking advantage of the increased mobile traffic as a consequence of the pandemic. YouTube, driven by mobile traffic growth saw also an important growth. And yet its ability to keep monetizing the traffic through increased ad spending, but also by serving more ads on its platforms. The above led to increased profitability, given the ability of Alphabet to reduce/or keep its costs stable as traffic increased. Google ads platform remains the strongest segment of the company, with massive profitability. However, both play a critical role in the future (Google Cloud Platform in the next five years, while the other bets in the coming ten-fifteen years). Google has negative margins on its Cloud Platform and its further bets. Facebook How much will Facebook (now Meta) lose as a consequence of Apple's policy change? At least $10 billion in 2022! This is for Facebook alone. For some context, Apple now requires apps to explicitly ask users if they want to be tracked, while before the consent was given by default. As Facebook highlighted in its latest financials: And they further explained: And finally, they put a number on it: "with Apple's iOS changes and new regulation in Europe, there's a clear trend where less data is available to deliver personalized ads... ...Like others in our industry, we faced headwinds as a result of Apple iOS changes... ...Apple created two challenges for advertisers: one is that the accuracy of our ads targeting decreased, which increased the cost of driving outcomes; the other is that measuring those outcomes became more difficult. These challenges are complex and interrelated. We're working to try and improve things..." "For example, by making progress in closing the underreporting gap for iOS web conversions and by introducing tools like our aggregated event measurement solutions to deliver better insights for advertisers..." "And, you know, we believe the impact of iOS overall as a headwind on our business in 2022 is on the order of $10 billion, so it's a pretty significant headwind for our business. And, you know, we're seeing that impact, you know, in a number of verticals..." Summing things up for the digital advertising landscape: Some key highlights from the latest financials of Google (Alphabet) and Facebook (Meta) also show some of the changes in the digital advertising landscape. This, in turn, is affecting millions of small businesses across the world: In reality, it was the pandemic that accelerated the trend. For both Google and Facebook, mobile traffic is the only growth driver. The pandemic determined a mobile transition: while companies like Google and Facebook have been investing in mobile-first for years. An interesting note here: Google (thanks to its vertical integration) managed to monetize its increased adoption. Facebook, on the contrary, hasn’t, and this was definitely due to Apple’s privacy change. The supply chain of data: it’s crucial to understand how Google and Facebook are entirely different in terms of structure. Where . While this played out well for Facebook so far. As soon as Apple tied up the distribution pipeline, it threatened the whole business! Google is vertically integrated (Hardware/Android > Chrome/Play Store > Search Engine/Search Ads), Facebook primarily relies on its strong brands who benefit from stickiness (Facebook, Instagram, WhatsApp, Messenger, Oculus) for the first time users stalled in various regions across the world. This is because there is too much pressure from new social media platforms (TikTok is killing it). Stalled user growth of Facebook: where Google’s growth potential has not been affected so far (it might come in the future), Facebook’s growth has already been affected. Thus, the move to the Metaverse for Facebook is not just a strategic bet; it’s a survival anchor! This means Facebook has to rush into it! Blitzscale the adoption of VR devices, which might turn into enormous mistakes. The move into the Metaverse becomes a survival anchor for Facebook: Who might benefit? Let’s draw some key points: Facebook has to push into VR (Oculus is a first step there). But the vision of Facebook of VR is way more pervading than any other company as its survival depends on that. Accelerating VR adoption for Facebook: all tech giants are rushing into VR with different philosophies. Apple believes in augmented reality, Google believes in mixed reality. Facebook wants to bring us into the Matrix. Therefore, we’ll see the first VR headsets probably in 2022 as the first exploration toward AR. VR/AR wars coming in 2022: digital platforms that go back to contextual advertising rather than aggressive targeting might benefit in this phase (Reddit over Facebook). Contextual advertising might win over targeted ads based on tracking: companies whose business model has privacy built-in will keep their growth trajectory. Thus, search engines like DuckDuckGo might skyrocket, and social media platforms like MeWe might thrive. Rise of privacy first search engines and social media: What about Amazon? Some interesting takes from Amazon's financials as it closed the year. No wonder Jeff Bezos picked up Andy Jessy (the current CEO) as his successor. Andy Jessy has led AWS since its inception in 2003! In Q4, Amazon online stores stalled compared to the previous year (hint: the pandemic contributed to the 2020 boost). The fastest-growing segments were AWS and Amazon advertising. Amazon's advertising machine passed $30 billion in revenues in 2021. For some context, YouTube ads generated over $28 billion in 2021. Amazon AWS showed incredible growth. A 40% growth year over year. Amazon AWS is a $71 billion revenue run rate business! And it has wide margins (so far) and incredible profitability, which connects to the next point. Of Amazon's over $33 billion in profits, over $18 billion were contributed by AWS! This means that more than 50% of Amazon's profitability depends on AWS. AWS is an enterprise powerhouse. Among its customers, there is also Meta! Meta, former Facebook Inc, relies on Amazon AWS for its AI development platform. Other huge brands built on top of Amazon AWS are Best Buy, Under Armour, Goldman Sachs, Pfizer, Roche, and many others. Can you guess what's next? Also Published Here

Big Tech Highlights From Alphabet, Meta and Amazon Financials For 2021

Fun little experiment to try out: go to your browser’s address bar (works best in desktop Chrome) and record your top websites that autocomplete for each of the letters . And if you’re feeling especially brave, share the results with the world… 😈 a-z My Google Chrome autocomplete results for A-Z. My Results A) - no surprises here. 🛍 amazon.com B) - mirror for the torrents. baypirateproxy.org thepiratebay C) - for all your anime needs. crunchyroll.com D) - kinda old school, but I prefer checking it to any personalized newesfeeds for generic news. digg.com E) - Useful TODO list app (I’ve recently switched to and couldn’t be happier). en.todoist.com todoist notion F) - Duh, though I’ve disabled my FB news feed with the chrome plugin so I don’t visit nearly as often as I used to. facebook.com News Feed Eradicator G) - Open source work consumes a lot of my time. ⭐️ is a reasonably distant second place here. github.com grubhub.com H) - Netflix’s younger cousin.. cause some stuff’s only available on Hulu. hulu.com I) - Really interesting new design tool that aims to make design + dev collaboration much easier. interplayapp.com J) - Very useful curated list of high quality React components. js.coach/react K) - Old KickAssTorrents mirror that is out of date. katcr.co L) - Ahh the joys of local web development. 😄 http://localhost:3000 M) - I use FB Messenger a lot more than SMS ever since I spent a few years living abroad in Thailand. International SMS is clunky but internet messaging platforms always “just work”. And of those platforms like Line, Wechat, etc., the only one where 99% of my friends have accounts is FB Messenger. messenger.com N) - Hacker News! 😎 I was a little surprised didn’t make it as my #1 but these days I stream more from external devices than my primary laptop. Also rounding out my top 3. news.ycombinator.com Netflix nytimes.com O) - Cause some companies force you to use Outlook 🤷‍ outlook.com P) - As a serial entrepreneur, Product Hunt’s a daily visit for me. producthunt.com Q) Leaving this one out as it’s an admin dashboard for async jobs for a recent project of mine. R) - The constant rabbit-hole of Reddit. :sigh: reddit.com S) - Torrent directory listing. seventorrentsmirror.info T) - This one’s probably my most visited. FOMO for better or worse… twitter.com U) - Excellent royalty-free stock photo site. unsplash.com V) - Crunchyroll’s cooler sister made by the same peeps. Anime streaming central. vrv.co W) - Reasonably unbiased news. wsj.com X) This was the only letter that didn’t autocomplete to any website for me… no seriously, it didn’t autocomplete even though there were results… I wonder if this is a built-in privacy mechanism of Chrome because I definitely don’t use Incognito mode... 🤷‍ Y) - I spend a lot more time on here than I probably should. :sigh: youtube.com Z) - Cause health’s important yo. zocdoc.com Categories These results can be roughly grouped into the following categories: Streaming (Hulu, VRV, YouTube, Crunchyroll, and various torrent sites) Software (localhost, GitHub, ProductHunt, js.coach) News (WSJ, Digg, Twitter, Hacker News) Social (Twitter, Facebook, Messenger) Insights I stream a LOT of video and ❤️ torrents (shout out to ). WebTorrent I’m pretty surprised there’s nothing NSFW on this list. 🕵 I’m sure there are a few results on here that would be novel and/or surprising to you and vice-versa. Conclusion So what is your alphabet of most visited websites? If you’re feeling brave, share some standouts from your list in the comments below or mention me @transitive_bs on Twitter! Hopefully this was a fun little experiment for you.. in a similar vein to ProductHunt sharing the most sticky apps that people use a lot but most people don’t know about ( ). example Also, gimme some of them claps 👏 👏 👏 Cheers!

🔥 Alphabet of Your Most Visited Sites 🔥

A is for because it’s traffic we cite, although I don’t know how accurate it is. We could google it. ALEXA.com ( ranking in parenthesis ) B is for because they power the majority of China’s 4b monthly search queries, and that’s just the beginning of their business. BAIDU.com (4th) C is for because the news is the news of the owner’s interest and seeing a site for what it is isn’t popular. CNN.com (89th) D is for because digital storage will forever be a challenge and starting a site with the “duh” sound isn’t popular. DROPBOX.com (72nd) E is for because they convinced the average citizen they could make money on the internet. They used to rank much higher. EBAY.com (32nd) F is for because it’s the back-hole that controls a billion identities on the internet. FACEBOOK.com (3rd) G is for because our clicks on it fund the Alphabet, and it’s now the verb we trust to end any argument. GOOGLE.com (1st) H is for because I am evaluating pros and cons to “likeminded sponsorship” in the age of internet journalism. HACKERNOON .com ( 6,358th ) I is for because it filters the mundane and they take no shame in the . INSTAGRAM.com (16th) stories of stealing digital products J is for because why not connect China with the world’s consumers. So few sites start with “j.” JD.com (337th) K is for because it’s a Russian movie site, and the only .club to be #1 ranked site for it’s letter. KINOGO.club (256th) L is for because we like to pretend that doesn’t have a monopoly on googling things. LIVE.com (12th) GOOGLE M is for because it ranks higher than it’s parent but lower than its brother MSN.com ( 42nd ) MICROSOFT (44th) (LIVE, 12th). N is for because they set the market rate for the cost of streaming . NETFLIX.com (50th) entertainment O is for because we mustn’t’ forget that eyeballs are made to let our fingers know which ads to click on. ONCLICKADS.com (31st) P is for because if there’s a better way to organize images, it’ll be acquired into this alphabetical world. PINTEREST.com (45th) Q is for because who doesn’t love consuming their Chinese news via endless rows of font 8 headlines with no pictures? QQ.com (10th) R is for because no site structure was ever more perfect for the specific nature of the internet than upvote/downvote. REDDIT.com (25th) S is for because wow another Chinese media site… No matter the government, traffic monetizes. SINA.com (19th) T is for because maybe someday I’ll write a letter home saying, “Hey Mom, you won’t believe it, they verified !!” TWITTER.com (8th) me U is for because Portuguese is a beautiful language and this is the largest internet portal for Portuguese speakers. UOL.com.br (126th) V is for because Facebook does not own Europe! This St. Petersburg based social network has 369,000,000+ users. VK.com (15th) W is for because it’s re-writing any term beyond the dictionary. It’s .com redirects to it’s .org, only one on the list! WIKIPEDIA.org (5th) X is for because porn is the driving factor behind much of the internet’s innovations. XVIDEOS.com(74th) Y is for because all video in the history of the world should be on one site owned by , right? YOUTUBE (2nd) GOOGLE Z is for because questions and answers are the light and dark of the internet. Other than the URL, this site is not in English. ZHIHU.com (135th) Editor’s Note: Each site is the top Alexa ranked for the letter it starts with, except for and because the internet! “A” “H” “AH,”

Alphabetical Internet Ownership

There are situations, when during the development of an application, there is a need to create a sequence of data, such as a sequence of numbers that can serve as identifiers. Some programming languages have special built-in functions that can generate such data, but JavaScript does not. range In this article, we will look at how to implement a similar function in JavaScript and how to use it in different situations. The function in JavaScript can be useful in cases where you need to generate a sequence of integers or other data types with a certain amount of spacing between values. Also, as an example, you can give a sequence of dates, with an interval of a day, a week, or a month. range However, let’s begin with an easier example and use the function to generate a sequence of numbers from 1 to 10 like this: range range(1, 10); // [1, 2, 3, ..., 10] Below you can see one of the implementations of the range function to generate a sequence of numbers: function range(start, end) {
  const seq = [];

  for (let i = start; i <= end; i++) {
    seq.push(i);
  }

  return seq;
} In this example above, the function creates a new array and fills it with integers using a loop. After the array is filled, the function returns it. range seq for You can change the step between elements of a sequence by adding a third argument to the function: function range(start, end, step) {
  const seq = [];

  for (let i = start; i <= end; i += step) {
    seq.push(i);
  }

  return seq;
}

range(1, 10, 2); // [1, 3, 5, 7, 9] Here, the function returns an array of odd numbers from 1 to 10 since the initial value is set to 1 and the step is set to 2. You can also change the direction of the sequence generation by specifying a negative step like this: range function range(start, end, step) {
  const seq = [];

  for (let i = start; i >= end; i += step) {
    seq.push(i);
  }

  return seq;
}

range(10, 1, -2); // [10, 8, 6, 4, 2] Note that the function only works with integers. If you need to generate a sequence of fractional numbers, then you need to write an appropriate function or use other ways to generate a sequence with fractional numbers. range The function can be used to create a sequence of strings, such as a sequence of letters from the alphabet: range function range(start, end, step = 1) {
  const start = start.charCodeAt(0);
  const end = end.charCodeAt(0);

  const seq = [];
  for (let i = start; i <= end; i += step) {
    seq.push(String.fromCharCode(i));
  }

  return seq;
}

range('a', 'z'); // ['a', 'b', 'c', ..., 'z'] In this example, we initiate the method on the start and end value of the sequence. The method returns the numeric value for the character at the specified index. charCodeAt charCodeAt Next, using the loop, iterate over the elements of the sequence and set the values using the method. The static method returns a string created from the specified character code. After the array is filled, the function returns it. for String.fromCharCode() String.fromCharCode() In the previous examples, the function generates either a sequence of integers or strings from start to end values with a given step. However, you can also create a function with more complex logic that will work on different data types at the same time. range range To summarize, we have analyzed the function in JavaScript, which allows you to generate a sequence of numbers and strings with a certain step. range We learned how to implement such a function with a loop and how to use it in various situations. Use the function to simplify your code and speed up development. for range

Function Range in JavaScript: Generating a Sequence of Data

The Markup obtained internal documents that coach new employees to avoid creating “very real legal risks” in using words like “market” and “networkeffects” As Google faces at least four major antitrust investigations on two continents, internal documents obtained by The Markup show its parent company, Alphabet, has been preparing for this moment for years, telling employees across the massive enterprise that certain language is off limits in all written communications, no matter how casual. The taboo words include “market,” “barriers to entry,” and “network effects,” which is when products such as social networks become more valuable as more people use them. “Words matter. Especially in antitrust law,” reads one document titled “ .” Five Rules of Thumb for Written Communications “Alphabet gets sued a lot, and we have our fair share of regulatory investigations,” reads another. “Assume every document will become public.” The internal documents appear to be part of a self-guided training session for a wide range of the company’s more than 100,000 employees, from engineers to salespeople. One document, titled “Global Competition Policy,” says it applies not only to interns and employees but also to temps, vendors, and contractors. The documents explain the basics of antitrust law and caution against loose talk that could have implications for government regulators or private lawsuits. "Words matter. Especially in antitrust law." Alphabet internal training document In one of the documents, which appear to be written by the legal team, employees are advised to choose their words carefully and use only third-party data when referencing Google’s “position in search” in sales pitches. They are further cautioned never to print or hand out their slides. “We use the term ‘User Preference for Google Search’ and never the term market share,” says. that document Google Search is the company’s most profitable product and, as such, a large target for antitrust regulation. It’s that nine in 10 web searches in the U.S. are completed on Google. estimated To take action against a company, antitrust regulators must establish that it has a dominant share in a market. The more broadly a market is defined, the easier it is for the company to argue that it has real competition. In the slides, employees are cautioned that defining a market is hard and best avoided. “These are completely standard competition law compliance trainings that most large companies provide to their employees,” Google spokesperson Julie Tarallo McAlister said in an email. “We instruct employees to compete fairly and build great products, rather than focus or opine on competitors. We’ve had these trainings in place for well over a decade.” Bad Words One part of the presentation, subtitled “ ,” advises employees on which terms are “Bad” and “Good.” Communicating Safely Instead of “market,” employees may say “industry,” “space,” “area,” or simply cite the region, according to the presentation. Instead of “network effects,” the presentation suggests “valuable to users.” And instead of “barriers to entry,” substitute “challenges.” Alphabet internal training document An excerpt from Alphabet internal training documents. Re-created by The Markup to protect sources. Alphabet is under investigation by 50 attorneys general and the Department of Justice for potentially abusing its dominance to undermine competition. Its acquisitions, along with those of other major U.S. tech giants, are being reviewed for anticompetitive effects by the Federal Trade Commission. The European Commission an “in-depth” investigation of Google’s acquisition of fitness tracker Fitbit and is probing possible antitrust violations regarding Google for Jobs. announced McAlister said Google representatives “continue to engage with ongoing investigations” and pointed to a where the company said its deal with Fitbit “will increase choice.” blog post She also said the company “vigorously” denies claims by Sonos in its patent infringement suit, which alleges that Google first partnered with the company and then copied its technology. Sonos CEO Patrick Spence in a hearing before the House Judiciary Committee in January that Google uses its dominance to shut down competition. Google countersued in June, alleging that Sonos infringed on its patents. testified Last week, Alphabet CEO Sundar Pichai faced a grueling 61 questions from lawmakers referencing the company’s dominance in online video, advertising, search, and other areas at a before the House Judiciary antitrust subcommittee. hearing During the hearing, the chair of the subcommittee, Rep. David Cicilline, mentioned an by The Markup that showed Google Search presented its own results and products above competitors’. He held it up as evidence that Google has turned search into a “walled garden.” investigation According to a CSPAN transcript of the hearing, Pichai did not say the words “network effects” or “dominant” once and steered clear of defining any particular market. “Google operates in highly competitive and dynamic global markets, in which prices are free or falling, and products are constantly improving,” he said. Why "Network Effects" Matter Respected Silicon Valley antitrust lawyers Doug Melamed of Stanford University and Gary Reback of Carr & Ferrell both said it’s not unusual for a giant company such as Alphabet to coach employees to avoid talking about dominance. “It’s not just trying to hide the truth, it’s really trying to avoid the use of inflammatory language,” said Melamed, a former acting assistant attorney general in charge of the antitrust division at the U.S. Department of Justice. “The use of the word ‘market’ could be very innocent, but it could be misleading and provocative to an antitrust enforcer. Antitrust enforcers really get exercised when they think someone dominates a market.” Alphabet’s emphasis in training documents on avoiding the term “network effects” is a new one, both lawyers said, and goes to one of the main issues in tech antitrust. It was one of the concepts the U.S. government and attorneys general from 20 states and the District of Columbia used to prosecute Microsoft for antitrust violations in 1998. They claimed the company broke the law, in part by using its dominance in the PC market to restrict competition by bundling Internet Explorer and other Microsoft products with the Windows operating system. “The barriers that exist to the entry of new competitors or the expansion of smaller existing competitors, including network effects, mean that dominance once achieved cannot readily be reversed,” the DOJ wrote in its complaint. “Once you admit strong network effects,” said Reback, who has represented competitors in antitrust cases against Google, “then it becomes easier for someone like me to say, ‘You’ve gotten this dominant share through exploitation of network effects, it’s clear there are high barriers to entry, we’re going to need to break you up or do something strong to stop you from abusing your position.’ ” "We are not out to ‘crush,’ ‘kill,’ ‘hurt,’ ‘block,’ or do anything else that might be perceived as evil or unfair." Alphabet training slide Reback represented Netscape early on in the Microsoft case and was the co-author of a about Microsoft that argued that network effects made tech companies gain market share faster than in other industries, and therefore antitrust regulators must be more aggressive. 222-page white paper Alphabet’s training documents lay out the basic concepts of antitrust law: “Competition laws, aka antitrust laws, are those that govern the conduct of businesses in order to protect fair competition, and, as a result, protect consumers.” No Crushing The slides also remind employees to keep the focus on how whatever they’re talking about benefits the public, which is a strong defense in antitrust matters: “Pro-tip: when drafting slides, always include at least one that shows how your business decisions will benefit consumers, partners, etc.” “We are not out to ‘crush,’ ‘kill,’ ‘hurt,’ ‘block,’ or do anything else that might be perceived as evil or unfair,” reads one portion. “Microsoft famously got into trouble when one of their employees threatened to ‘cut off Netscape’s air supply.’ ” Employees are given “Five Rules of Thumb for Written Communications,” according to the documents: We’re out to help users, not hurt competitors. Our users should always be free to switch, and we don’t lock anyone in. We’ve got lots of competitors, so don’t assume we control or dominate any market. Don’t try and define a market or estimate our market share. Assume every document you generate, including email, will be seen by regulators. Alphabet internal training document s. An excerpt from Alphabet internal training documents. Re-created by The Markup to protect source “If they can get their employees to comply with those guidelines, good for them,” said Melamed, the former DOJ antitrust official, who was one of the prosecutors on the Microsoft case. “But most companies’ documents are still filled with all sorts of stuff like this.” European regulators have already used internal emails against Alphabet. In an earlier investigation, they found emails that indicated that its shopping product’s web traffic doubled in 2008, mostly because it was featured in search listings. The European Commission concluded that the company had abused its dominance and in 2017 €2.42 billion, a record at the time. “Those emails were quoted selectively and out of context,” McAlister said. fined the company Reback, the antitrust lawyer, said there is a limit to how internal communications can be used, especially when they’re written by lower-level employees. In general, anticompetitive intent is not part of the legal argument, he said. Rather, lawyers for the plaintiff or the government must establish an anticompetitive effect. "I didn’t mean to imply that we’d be buying them to prevent them from competing with us in any way.…" Facebook CEO Mark Zuckerberg Internal communications from higher-level executives that show apparent anticompetitive intent may have more weight—or at least, generate more publicity. The House Judiciary antitrust subcommittee, which held last week’s hearing, collected text messages in which a Facebook executive suggested that CEO Mark Zuckerberg would go into “destroy mode” if Instagram turned down an acquisition offer. The committee also released an email Zuckerberg wrote musing that “there are network effects around social products” and buying companies like Instagram “will give us a year or more to integrate their dynamics before anyone can get close to their scale again.” He sent another email less than an hour later that, “I didn’t mean to imply that we’d be buying them to prevent them from competing with us in any way.… I’m mostly excited about what the companies could do together if we worked to build what they’ve invented into more people’s experiences.” clarifying During its 2012 antitrust investigation into Google, the FTC also collected emails, some of which Alphabet’s legal team would have frowned at. In one, Google’s top economist, Hal Varian, celebrated that market research firm Comscore had undervalued Google’s market share. “From an antitrust perspective,” Varian wrote, “I’m happy to see them underestimate our share.” The FTC ultimately overrode staff recommendations to take legal action against Google and closed the investigation without filing charges. Credits Originally published as " " with the Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) license. To Head Off Regulators, Google Makes Certain Words Taboo

Can You Speak The Google Language, Please? 

Comment comes as part of big tech CEO grilling by antitrustcommittee In a historic hearing questioning the heads of four of the nation’s five largest technology companies, the chair of Congress’s antitrust subcommittee, David N. Cicilline, on Wednesday directed sharp criticism at Sundar Pichai, CEO of Google’s parent company Alphabet, pointing to The Markup’s findings that Google Search places the company’s products and results above competitors’. “An investigative report published just yesterday found that 63 percent of searches that start on Google also end on Google’s own websites,” Cicilline said, referring to our findings about results on the coveted first screen. “And to me that is evidence that Google is increasingly a walled garden which keeps users on Google’s sites, even if Google doesn’t have the most relevant information. And it’s economically catastrophic for other companies online.” Cicilline mentioned the report after asking, “Increasingly what Google shows is what is most profitable for Google. Isn’t there a fundamental conflict of interest?” Pichai responded, “In a vast majority of queries on Google, we don’t show ads at all.” The hearing was the sixth in a year-long series titled Online Platforms and Market Power, which has included testimony from experts, competitors, and lower-level executives from the four tech giants, Google, Apple, Facebook, and Amazon. It was the first time all four had been questioned together, and Amazon CEO Jeff Bezos’s first appearance before Congress. They were asked about their respective market dominance and treatment of competitors. And some members of the committee asked the CEOs about their treatment of conservative media and causes. The hearing came one day after a months-long, painstaking investigation into Google Search that for the first time quantified how much the company highlights its own products and serves “answers” on the results page, many of which take information from other websites and are meant to satisfy queries without the need to leave the Google universe of products. The Markup published Our of more than 15,000 U.S. trending queries found that Google gave 41 percent of the first search results page to itself and dominated the coveted first screen. sophisticated analysis Google spokesperson Lara Levin called The Markup’s “flawed and misleading.” She criticized it for being “based on a non-representative sample of searches” and said using trending searches makes it more likely that results would include Google “knowledge panels”—information summaries curated from various sources—than a random sample would. As we wrote earlier, Google has not publicly released a random sample of searches. methodology Among the competitors who told The Markup that Google’s decision to highlight its own products in search is hurting other businesses is Chris Cummings, CEO of Curiosity Media, which owns SpanishDict.com. Cummings said that when Google serves SpanishDict as the top result, the site gets 80 percent of clicks—but when Google Translate appears above it, that figure drops to 2 percent. Google’s decisions have limited the company’s growth, he said: “We only exist because there are still some queries where they don’t put their stuff at the top.” Travel websites large and small said boosting Google Flights and Google Hotels on Google’s search results page has taken away business, even though some pay Google billions for ads. To determine the amount of space Google dedicates to itself, we built a custom scraper to gather U.S. Google Trends for two months, starting in November 2019. We built another scraper to run the searches as they would appear on mobile devices, where the majority of searches now occur. We wrote more than 1,000 lines of code to parse and analyze the resulting dataset, which contained more than one million rows. Members of Congress questioned Pichai, Bezos, Apple CEO Tim Cook, and Facebook CEO Mark Zuckerberg in a hearing that started more than an hour late and was ongoing as this story was published. During Rep. Greg Steube’s time for questions, he accused Google company Gmail of censoring Republican email. The Markup that Gmail was putting advocacy and political email from across the spectrum in its marketing tab rather than delivering it in the main inbox. (We found no political bias.) reported in February Pichai replied to Steube that the company is doing what users want: segregating bulk emails from messages from friends and family. Credits Originally published as " Congressman Says The Markup Investigation Proves Google Has Created a “Walled Garden” " with the Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) license.

The Markup Investigation: Congressman Says that Google engineered a “Walled Garden”

Heavy is the head that wears the crown during a Wall Street bull run, and Nvidia’s status as the king of the ongoing generative AI boom means that the stock can be especially vulnerable to the prospect of weakening investor appetite. Nvidia (NASDAQ:NVDA) became Wall Street’s most exciting story in 2023. The chipmaker rallied some 239% throughout the year, smashing through the $1 trillion market capitalization glass ceiling before peaking at a cap in excess of $2tn. The optimism, fuelled by the ongoing generative AI market rally, continued to make waves in the first-quarter of 2024. NVDA rallied almost 90% in the opening three months of the year to continue its strong spell of dominance. But finally, at the beginning of April, the stock took its biggest tumble in four years. The wider implications of the Nvidia sell-off are likely to shake up the entire generative AI market. Counting the Cost of NVDA Sell-Offs The first three weeks of April 2024 saw NVDA shed more than 15% of its value amid widespread investor sell-offs. Although greater degrees of volatility can be expected for stocks caught up in the hype phase of an emerging market, Nvidia’s April downturn has been the company’s biggest dip since the initial outbreak of the Covid-19 pandemic in March 2020. For Nvidia, the pandemic era must seem like a lifetime ago. Its March 2020 sell-offs saw the stock tumble to $51.44. The stock’s peak closing price in March 2024 was $950.02. But after four years of meteoric growth, Nvidia’s recent market downturn is particularly concerning–particularly now the stock’s gained the status as a tech pioneer that’s been driving the S&P 500 higher on a consistent basis. Nvidia’s position at the pinnacle of the generative AI boom could mean that the stock would always be the first to absorb the impact of a change in investor sentiment, and the fears are that NVDA would be the canary in the mine should we suddenly discover that generative AI has been in a hype bubble that’s ready to burst for some time. Although Nvidia’s recent difficulties have coincided with the news that US inflation rates have continued to defy expectations, placing the prospect of interest rate cuts on hold for the foreseeable future, analysts suggest that fears over a cooling generative AI boom remain the key cause of investor concern in the wake of the underperformance. “The stock pullback has very little to do with rates,” explained Parag Thatte, strategist at Deutsche Bank, in a recent Financial Times interview. “It’s more to do with investors pricing in slower earnings growth.” There’s also a growing concern that generative AI has begun to transform from its hype phase towards the question of implementation. While Large Language Models (LLMs) like ChatGPT have been excellent tools in showcasing the sheer power and potential of generative AI, they’ve also been largely error-prone in a manner that could cause more hesitancy among enterprises seeking to embrace the technology. It’s for this reason that Gil Luria, analyst at DA Davidson, warned that a 20% slide in Nvidia stock could take place by the end of 2024. Luria suggests that the implementation phase of Nvidia’s products, which are currently being bought en masse by the likes of Microsoft and Amazon, could run into trouble when key buyers reach capacity and demand begins to fall. However, other analysts suggest that the end of the generative AI hype phase will be met by a more strategic implementation phase which could keep the industry growing. “Concerns about GenAI failing to meet expectations in 2024 should be seen in the context of the natural cycle of hype and maturity of new technologies,” said Maxim Manturov, head of investment research at Freedom Finance Europe. “While some industry leaders have expressed frustration with the speed of progress in artificial intelligence and GenAI in their organisations, others remain optimistic about its potential.” “Companies heavily invested in GenAI are under pressure to prove profitability, and industry reports indicate a shift to more strategic AI initiatives in the coming years,” Manturov added. Is the Generative AI an Exceptional Case? Here, it’s worth reminding ourselves that we’re talking about the generative AI boom. This is an industry that Bloomberg Intelligence last year suggested would grow at a CAGR of 42% into a $1.3 trillion industry by 2032. Fresh demand for generative AI products was predicted to add around $280 million in new software revenue. Could we be talking about an emerging technology that’s market impact is without precedent? Morgan Stanley analysts have underlined the indecisive outlook for Nvidia and the wider generative AI market in Q2 2024 by proposing a more bullish outlook for NVDA as well as the generative AI-focused Amazon (AMZN). As Nvidia reported Q4 2023 revenue increases of 265% to $22 billion, CEO Jensen Huang claimed that "accelerated computing and generative AI have hit the tipping point," before adding that "demand is surging worldwide across companies, industries and nations." The company’s innovations in recent weeks have included the text-to-3D tool, LATTE3D, which Nvidia described as a ‘virtual 3D printer’, as well as a collaboration with Hipocratic AI to develop a network of generative AI nurses that could cost $9 an hour. Nvidia’s quick 5% rebound following its record sell-offs should serve as a wider vote of confidence in the innovation that the company is capable of. Regardless of whether its success has made it the GenAI canary, NVDA is certainly an innovative stock that’s capable of sustaining its impressive growth. Bubble or Boom? Ultimately, Nvidia’s performance over the past 12 months means that the stock has become a live litmus test for investor sentiment in generative AI. Whether that’s a good thing or not depends wholly on the capabilities of the industry itself. Nvidia’s rapid growth has seen the stock rub shoulders with the likes of Microsoft, Apple, and Alphabet, and the company will be keen to follow the lead of other industry giants to continue sustaining its growth through innovation. As a semiconductor leader, the possibilities are nearly endless for Nvidia. As long as the capabilities of generative AI continue to astound clients and deliver on its promises, NVDA remains a stock that can continue to grow. If investors get a whiff of saturation or inefficiency, the Nvidia litmus test could face some further challenges ahead. Heavy is the head that wears the crown during a Wall Street bull run, and Nvidia’s status as the king of the ongoing generative AI boom means that the stock can be especially vulnerable to the prospect of weakening investor appetite. Nvidia (NASDAQ:NVDA) became Wall Street’s most exciting story in 2023. The chipmaker rallied some 239% throughout the year, smashing through the $1 trillion market capitalization glass ceiling before peaking at a cap in excess of $2tn. Nvidia The optimism, fuelled by the ongoing generative AI market rally, continued to make waves in the first-quarter of 2024. NVDA rallied almost 90% in the opening three months of the year to continue its strong spell of dominance. But finally, at the beginning of April, the stock took its biggest tumble in four years. The wider implications of the Nvidia sell-off are likely to shake up the entire generative AI market. Counting the Cost of NVDA Sell-Offs The first three weeks of April 2024 saw NVDA shed more than 15% of its value amid widespread investor sell-offs. Although greater degrees of volatility can be expected for stocks caught up in the hype phase of an emerging market, Nvidia’s April downturn has been the company’s biggest dip since the initial outbreak of the Covid-19 pandemic in March 2020. For Nvidia, the pandemic era must seem like a lifetime ago. Its March 2020 sell-offs saw the stock tumble to $51.44. The stock’s peak closing price in March 2024 was $950.02. But after four years of meteoric growth, Nvidia’s recent market downturn is particularly concerning–particularly now the stock’s gained the status as a tech pioneer that’s been driving the S&P 500 higher on a consistent basis. Nvidia’s position at the pinnacle of the generative AI boom could mean that the stock would always be the first to absorb the impact of a change in investor sentiment, and the fears are that NVDA would be the canary in the mine should we suddenly discover that generative AI has been in a hype bubble that’s ready to burst for some time. Although Nvidia’s recent difficulties have coincided with the news that US inflation rates have continued to defy expectations, placing the prospect of interest rate cuts on hold for the foreseeable future, analysts suggest that fears over a cooling generative AI boom remain the key cause of investor concern in the wake of the underperformance. prospect of interest rate cuts on hold prospect of interest rate cuts on hold “The stock pullback has very little to do with rates,” explained Parag Thatte , strategist at Deutsche Bank, in a recent Financial Times interview. “It’s more to do with investors pricing in slower earnings growth.” explained Parag Thatte explained Parag Thatte There’s also a growing concern that generative AI has begun to transform from its hype phase towards the question of implementation. While Large Language Models (LLMs) like ChatGPT have been excellent tools in showcasing the sheer power and potential of generative AI, they’ve also been largely error-prone in a manner that could cause more hesitancy among enterprises seeking to embrace the technology. largely error-prone largely error-prone It’s for this reason that Gil Luria, analyst at DA Davidson, warned that a 20% slide in Nvidia stock could take place by the end of 2024. Luria suggests that the implementation phase of Nvidia’s products, which are currently being bought en masse by the likes of Microsoft and Amazon, could run into trouble when key buyers reach capacity and demand begins to fall. 20% slide in Nvidia stock 20% slide in Nvidia stock However, other analysts suggest that the end of the generative AI hype phase will be met by a more strategic implementation phase which could keep the industry growing. “Concerns about GenAI failing to meet expectations in 2024 should be seen in the context of the natural cycle of hype and maturity of new technologies,” said Maxim Manturov, head of investment research at Freedom Finance Europe . “While some industry leaders have expressed frustration with the speed of progress in artificial intelligence and GenAI in their organisations, others remain optimistic about its potential.” Freedom Finance Europe Freedom Finance Europe “Companies heavily invested in GenAI are under pressure to prove profitability, and industry reports indicate a shift to more strategic AI initiatives in the coming years,” Manturov added. Is the Generative AI an Exceptional Case? Here, it’s worth reminding ourselves that we’re talking about the generative AI boom. This is an industry that Bloomberg Intelligence last year suggested would grow at a CAGR of 42% into a $1.3 trillion industry by 2032 . $1.3 trillion industry by 2032 $1.3 trillion industry by 2032 Fresh demand for generative AI products was predicted to add around $280 million in new software revenue. Could we be talking about an emerging technology that’s market impact is without precedent? Morgan Stanley analysts have underlined the indecisive outlook for Nvidia and the wider generative AI market in Q2 2024 by proposing a more bullish outlook for NVDA as well as the generative AI-focused Amazon (AMZN). more bullish outlook more bullish outlook As Nvidia reported Q4 2023 revenue increases of 265% to $22 billion, CEO Jensen Huang claimed that "accelerated computing and generative AI have hit the tipping point," before adding that "demand is surging worldwide across companies, industries and nations." claimed that claimed that The company’s innovations in recent weeks have included the text-to-3D tool , LATTE3D, which Nvidia described as a ‘virtual 3D printer’, as well as a collaboration with Hipocratic AI to develop a network of generative AI nurses that could cost $9 an hour. text-to-3D tool text-to-3D tool collaboration with Hipocratic AI collaboration with Hipocratic AI Nvidia’s quick 5% rebound following its record sell-offs should serve as a wider vote of confidence in the innovation that the company is capable of. Regardless of whether its success has made it the GenAI canary, NVDA is certainly an innovative stock that’s capable of sustaining its impressive growth. Bubble or Boom? Ultimately, Nvidia’s performance over the past 12 months means that the stock has become a live litmus test for investor sentiment in generative AI. Whether that’s a good thing or not depends wholly on the capabilities of the industry itself. Nvidia’s rapid growth has seen the stock rub shoulders with the likes of Microsoft, Apple, and Alphabet, and the company will be keen to follow the lead of other industry giants to continue sustaining its growth through innovation. As a semiconductor leader, the possibilities are nearly endless for Nvidia. As long as the capabilities of generative AI continue to astound clients and deliver on its promises, NVDA remains a stock that can continue to grow. If investors get a whiff of saturation or inefficiency, the Nvidia litmus test could face some further challenges ahead.

Nvidia’s Shaky Q2 Underlines the Stock as Generative AI’s Sentiment Litmus Test

এনভিডিয়ার শ্যাকি Q2 জেনারেটিভ এআই এর সেন্টিমেন্ট লিটমাস টেস্ট হিসাবে স্টককে আন্ডারলাইন করেছে

Nvidias wackeliges zweites Quartal unterstreicht die Aktie als Lackmustest für die Stimmung von Generative AI

El inestable segundo trimestre de Nvidia subraya la acción como prueba de fuego del sentimiento de la IA generativa

Le Shaky Q2 de Nvidia souligne le titre en tant que test décisif du sentiment de Generative AI

एनवीडिया की अस्थिर Q2 स्टॉक को जेनरेटिव एआई के सेंटीमेंट लिटमस टेस्ट के रूप में रेखांकित करती है

エヌビディアの第2四半期の不安定さは、ジェネレーティブAIの感情リトマス試験として株価を浮き彫りにする

Nvidia의 Shaky Q2는 Generative AI의 감정 리트머스 테스트로서 주식을 강조합니다.

Shaky Q2 da Nvidia destaca o estoque como teste decisivo de sentimento da IA generativa

Нестабильный второй квартал Nvidia подчеркивает, что акции являются лакмусовой бумажкой настроений генеративного ИИ

Nvidia'nın Titrek 2. Çeyreği Hisse Senedinin Üretken Yapay Zekanın Duyarlılık Turnusol Testi Olduğunu Vurguluyor

Q2 run rẩy của Nvidia nhấn mạnh cổ phiếu là bài kiểm tra tình cảm của AI sáng tạo

Nvidia 第二季度业绩表现不佳，凸显其股票是生成式 AI 情绪的试金石

The headline might be a slight exaggeration — I'm not Elon Musk. However, I am bootstrapping a venture studio called Evernomic with a focus on the media industry. Our portfolio includes several newsletters, media brands, SaaS projects, service-based startups, high-authority websites, external contracts, and more. Evernomic is also quite a unique organization when it comes to our team and how we allocate our time and resources. Our team is now comprised of over 20 people, with the oldest being 26 years old and the average age is just over 21 years old. Such a young team has its advantages, but it also comes with the drawbacks associated with the stereotypes often held against younger people in business. At the end of the day, we don't have that chronological experience behind us, which also applies to time management. In addition, one value we all share is despising a future of becoming one-dimensional. We don't just aspire to be successful at business or any other aspect of life. We exercise and take care of our health, we spend much of our time learning new things, we like to socialize and make memories, we pay attention to how we present ourselves, and we pursue hobbies such as writing and public speaking. Now, this begs the question of how we manage our time so as not to fall behind on our responsibilities. Maintenance and Expansion Framework I have developed what I call the Maintenance and Expansion Framework. I divide my workload and things that require headspace into two categories: Maintenance: ongoing commitments requiring my continued attention on a regular basis that are necessary for me to keep going in the same trajectory. In short, these are things I must do to maintain my current progress.
Expansion: one-time or open-ended tasks that improve the overall condition of my life, whether it's in business, fitness, or happiness. This may sound controversial to openly admit, but I avoid commitments, otherwise defined as Maintenance Tasks, at all costs. That's the danger zone. It takes away from my ability to expand and take on more things when I'm faced with better opportunities in the future. I always prefer to leave some free margin to try new things, waste some time, let things go wrong without detrimental consequences, and so on. Therefore, I'm very picky with the things I commit to, and I try to minimize them. Conversely, I'm unafraid of the Expansion side of things. I can take on several things simultaneously as long as I know it will be over soon. The only decisions that truly concern me are those that lock me into a certain future. A typical day for me includes my daily routine, immediately tackling my Maintenance Tasks and later tackling three or four Expansion Tasks at a time. This division gives a game-changing order to my life. I know what the bare minimum of what I should be doing is, and it lets me focus on multiple domains within my life and business simultaneously. Collaborative Portfolio In addition to how I personally approach time management, our strategy revolves around creating a system that goes "full circle." This means that every project we take on must contribute in one way or another to other projects in our portfolio as well. I wrote a post some time ago about how, with this approach, we technically have more than 24 hours in a day because every hour we spend on one project also leads to some sort of progress for our other projects. To give you an example, we are planning on launching an ad network in the following months. What's the biggest challenge with such a business? Finding sufficient creators and advertisers to make the marketplace functional — the chicken and egg problem. Looking at companies who have been most successful at growing an ad network, such as Alphabet or Meta, they have always had at least one side of the equation satisfied prior. In Meta's case, they already had access to a large audience who were using their free products before they began rolling out advertisements. Over a year prior to our scheduled launch date, we acquired FindNewsletters.com, an authority newsletter directory. We have yet to make a penny off of this acquisition. However, we now have direct contacts with most newsletter creators, which gives us an extensive creator network to start with. Our job is now much simpler; we only need to find advertisers. Within Evernomic, we try to build similar systems and frameworks through which we push our projects that give us a streamlined and efficient approach. Moreover, tasks get delegated effectively, which minimizes commitments to low-yield tasks among our more experienced members. Naturally, focusing on so many things simultaneously has its drawbacks, too. I often fail to match deadlines, I often don't end up acting on plans I've had, and I often don't reach my goals. In fact, at any given moment, I'm probably not doing too well. However, I've learned that time has a magic of its own. Whenever I've looked back at my progress over a few months, I was pleasantly surprised. Now, I trust the process and keep pushing forward regardless. The headline might be a slight exaggeration — I'm not Elon Musk. However, I am bootstrapping a venture studio called Evernomic with a focus on the media industry. Our portfolio includes several newsletters, media brands, SaaS projects, service-based startups, high-authority websites, external contracts, and more. Evernomic is also quite a unique organization when it comes to our team and how we allocate our time and resources. Our team is now comprised of over 20 people, with the oldest being 26 years old and the average age is just over 21 years old. Such a young team has its advantages, but it also comes with the drawbacks associated with the stereotypes often held against younger people in business. At the end of the day, we don't have that chronological experience behind us, which also applies to time management. In addition, one value we all share is despising a future of becoming one-dimensional. We don't just aspire to be successful at business or any other aspect of life. We exercise and take care of our health, we spend much of our time learning new things, we like to socialize and make memories, we pay attention to how we present ourselves, and we pursue hobbies such as writing and public speaking. Now, this begs the question of how we manage our time so as not to fall behind on our responsibilities. Maintenance and Expansion Framework I have developed what I call the Maintenance and Expansion Framework. I divide my workload and things that require headspace into two categories: Maintenance: ongoing commitments requiring my continued attention on a regular basis that are necessary for me to keep going in the same trajectory. In short, these are things I must do to maintain my current progress. Expansion: one-time or open-ended tasks that improve the overall condition of my life, whether it's in business, fitness, or happiness. Maintenance: ongoing commitments requiring my continued attention on a regular basis that are necessary for me to keep going in the same trajectory. In short, these are things I must do to maintain my current progress. Maintenance: Expansion: one-time or open-ended tasks that improve the overall condition of my life, whether it's in business, fitness, or happiness. Expansion: This may sound controversial to openly admit, but I avoid commitments, otherwise defined as Maintenance Tasks, at all costs. That's the danger zone. It takes away from my ability to expand and take on more things when I'm faced with better opportunities in the future. I always prefer to leave some free margin to try new things, waste some time, let things go wrong without detrimental consequences, and so on. Therefore, I'm very picky with the things I commit to, and I try to minimize them. Conversely, I'm unafraid of the Expansion side of things. I can take on several things simultaneously as long as I know it will be over soon. The only decisions that truly concern me are those that lock me into a certain future. A typical day for me includes my daily routine, immediately tackling my Maintenance Tasks and later tackling three or four Expansion Tasks at a time. This division gives a game-changing order to my life. I know what the bare minimum of what I should be doing is, and it lets me focus on multiple domains within my life and business simultaneously. Collaborative Portfolio In addition to how I personally approach time management, our strategy revolves around creating a system that goes "full circle." This means that every project we take on must contribute in one way or another to other projects in our portfolio as well. I wrote a post some time ago about how, with this approach, we technically have more than 24 hours in a day because every hour we spend on one project also leads to some sort of progress for our other projects. To give you an example, we are planning on launching an ad network in the following months. What's the biggest challenge with such a business? Finding sufficient creators and advertisers to make the marketplace functional — the chicken and egg problem. Looking at companies who have been most successful at growing an ad network, such as Alphabet or Meta, they have always had at least one side of the equation satisfied prior. In Meta's case, they already had access to a large audience who were using their free products before they began rolling out advertisements. Over a year prior to our scheduled launch date, we acquired FindNewsletters.com, an authority newsletter directory. We have yet to make a penny off of this acquisition. However, we now have direct contacts with most newsletter creators, which gives us an extensive creator network to start with. Our job is now much simpler; we only need to find advertisers. Within Evernomic, we try to build similar systems and frameworks through which we push our projects that give us a streamlined and efficient approach. Moreover, tasks get delegated effectively, which minimizes commitments to low-yield tasks among our more experienced members. Naturally, focusing on so many things simultaneously has its drawbacks, too. I often fail to match deadlines, I often don't end up acting on plans I've had, and I often don't reach my goals. In fact, at any given moment, I'm probably not doing too well. However, I've learned that time has a magic of its own. Whenever I've looked back at my progress over a few months, I was pleasantly surprised. Now, I trust the process and keep pushing forward regardless.

How I Manage Over a Dozen Startups

In programming, we often need to go between strings and bytes. Humans read strings. Computers read bytes. As a result, dotnet developers need to understand how to convert a string to a byte array in C#. We accomplish this through a process called encoding and decoding to go back and forth between the two. In this article, I’ll provide you code examples to convert a string to a byte array in C# — and back again! You’ll also learn about some of the nuances of character encodings to look out for! Encoding and Decoding in C# If we want to convert a string to a byte array — or go the other way — we need to understand the concept of encoding and decoding. In software engineering, encoding refers to the process of transforming a string into a sequence of bytes, while decoding involves the reverse process of transforming bytes back into a string. Simple, right? Remember that strings are a sequence of characters, and the concept of a character makes a lot of sense to us as readers, but computers understand bytes. Characters themselves can be represented by numbers in computers so when we want to work with strings at a lower level, such as when sending data over a network or storing it in a file, we need to convert the string into a byte array. Encoding comes into play when converting strings to byte arrays. It determines how the characters in the string are represented as bytes. The encoding scheme defines the mapping between characters and their byte representations. Common encoding schemes include UTF-8, UTF-16, ASCII, and Unicode. Choosing the correct encoding is important because different encoding schemes support different sets of characters. For example, ASCII only supports the basic English alphabet (with a few more characters), while UTF-8 and UTF-16 are capable of representing characters from multiple languages and scripts. More on this later though since I know you’re eager for some code! Choosing an Encoding for String to Byte Array Conversion When converting strings to byte arrays in C#, one of the most important things to consider is character encoding. Character encoding determines the mapping between characters and byte values, and if you’re considering transforming data one way then you may want to put some thought into how to transform it back! What’s that supposed to mean? Well, if we use a data transform — let’s say taking a string and transforming it to bytes with an ASCII encoding — if there is not a mapping of a particular character to a byte representation we lose that data in the result. Now what happens if you want to go the other way and get your byte array back to a string? Data is missing! In C#, there are various encodings available, including ASCII, UTF-8, and UTF-16, each with its specific characteristics and usage scenarios. Let’s explore these encodings and see how they can be used for string to byte array conversion. ASCII Encoding ASCII encoding represents characters using 7 bits, allowing for a total of 128 different characters. It’s primarily suitable for handling basic English characters and is more space-efficient compared to other encodings. Here’s an example of converting a string to a byte array using ASCII encoding: string text = "Hello, World!";
byte[] asciiBytes = Encoding.ASCII.GetBytes(text); In many modern applications, ASCII may not be what you’re after. This is especially true if you have users across the globe in locales that are not English. That’s not to say that ASCII *cannot* be used, but you’ll want to be careful about what data you ASCII encode so as to not lose information during the encoding transform. So keep in mind that the ASCII character set represents characters using a 7-bit encoding scheme, allowing for a total of 128 unique characters. However, with the growing need for internationalization and multilingual support, ASCII alone is insufficient to represent all characters effectively. Good thing we have some other options coming right up! UTF-8 Encoding UTF-8 encoding is a variable-length encoding scheme that can represent any Unicode character. It’s widely used for encoding text in various languages and is backward compatible with ASCII. It uses fewer bytes for representing ASCII characters but may require more bytes for non-ASCII characters. Here’s an example of converting a string to a byte array using UTF-8 encoding: string text = "Привет, мир!";
byte[] utf8Bytes = Encoding.UTF8.GetBytes(text); UTF-16 Encoding UTF-16 encoding represents characters using either 2 or 4 bytes, making it capable of representing any Unicode character. It’s commonly used by applications that need to handle multilingual text or when interoperability with other systems is required. The encoding can be little-endian or big-endian, with the former being more prevalent. Here’s an example of converting a string to a byte array using UTF-16 encoding: string text = "こんにちは、世界！";
byte[] utf16Bytes = Encoding.Unicode.GetBytes(text); UTF-8 Encoding vs UTF-16 Encoding – What’s The Difference? Each of these encodings is variable width, and their size gives them different characteristics when we consider different alphabets. Starting with UTF-8, it’s variable-width and backward-compatible with ASCII. In this encoding: ASCII characters, which are on the range U+0000 to U+007F, take only 1 byte.
In the next range, code points U+0080 to U+07FF take twice as much space at 2 bytes each.
Code points U+0800 to U+FFFF take one more byte, bringing us up to 3 bytes
And finally, code points U+10000 to U+10FFFF take 4 bytes. This can be very effective for English text because those characters will take up minimal space. However, when it comes to Asian text it’s not ideal for the exact opposite reason. Because UTF-16 has code points from U+0000 to U+FFFF take 2 bytes and code points U+10000 to U+10FFFF are double that at 4 bytes, it’s not-so-great for English. It *does* happen to be better suited for Asian characters though. There’s even UTF-32 encoding! This is a fixed-width encoding where all of the code points take four bytes — unlike the others I’ve mentioned which are dynamic in size. This can use much more storage than the other encodings, but because of its simplicity can be much faster to operate on. Give some consideration to the alphabets you’ll need to support primarily! Best Practices for String to Byte Array Conversion When converting strings to byte arrays in C#, it’s important to follow best practices to ensure efficiency and reliability. In this section, I’ll discuss some key best practices that you should keep in mind when performing string-to-byte array conversions. Error Handling and Validation When working with encodings, it’s important to handle potential errors and validate your data to prevent unexpected behavior in your code. Ideally, you structure the flow of your application such that you know what kind of data you’re dealing with. If you can write code to avoid errors in the first place, this is preferred! We don’t want to have to rely on this but sometimes it’s outside of our control — handling errors is by using try-catch blocks. By encapsulating the conversion code within a try block, you can catch any exceptions that may occur during the conversion process and handle them gracefully. If you don’t have control over the source of the input data, this is something you might need to do for safety. Encoding Selection C# provides several encoding options for converting strings to byte arrays, such as UTF-8, UTF-16, ASCII, and more. It’s important to select the appropriate encoding based on the specific requirements of your application. Consider factors such as character sets, compatibility with other systems, and performance implications when choosing the encoding. // Example of encoding selection
string inputString = "Hello, World!";
byte[] encodedBytes = Encoding.UTF8.GetBytes(inputString); Just like we saw in the previous examples, after the Encoding class we get to pick the static property holding the encoding instance. If we need to select one to pass around as a variable and parameter, you can absolutely store it in a dedicated encoding reference: Encoding selectedEncoding = Encoding.UTF8;
SomeMethod("Hello World!", selectedEncoding); Selecting the wrong encoding can have big consequences for your application! This is especially true if you save data with an encoding that will lose data resolution and you can’t reverse it… so put some care into this! Now You Know How to Convert a String to Byte Array in C#! You’re a pro now with encoding and decoding! Well, maybe not a full-on expert… but you have the basics put in front of you and some guidelines to work with. That’s a pretty good start. Remember to select the right encoding for the situation you’re dealing with. Keep in mind that you can potentially lose data resolution when using the wrong encoding, and as a result, lose data forever! If you found this useful and you’re looking for more learning opportunities, consider subscribing to my free weekly software engineering newsletter and check out my free videos on YouTube! In programming , we often need to go between strings and bytes. Humans read strings. Computers read bytes. As a result, dotnet developers need to understand how to convert a string to a byte array in C#. We accomplish this through a process called encoding and decoding to go back and forth between the two. programming In this article, I’ll provide you code examples to convert a string to a byte array in C# — and back again! You’ll also learn about some of the nuances of character encodings to look out for! Encoding and Decoding in C# If we want to convert a string to a byte array — or go the other way — we need to understand the concept of encoding and decoding. In software engineering, encoding refers to the process of transforming a string into a sequence of bytes, while decoding involves the reverse process of transforming bytes back into a string. Simple, right? Remember that strings are a sequence of characters, and the concept of a character makes a lot of sense to us as readers, but computers understand bytes. Characters themselves can be represented by numbers in computers so when we want to work with strings at a lower level, such as when sending data over a network or storing it in a file, we need to convert the string into a byte array. Encoding comes into play when converting strings to byte arrays. It determines how the characters in the string are represented as bytes. The encoding scheme defines the mapping between characters and their byte representations. Common encoding schemes include UTF-8, UTF-16, ASCII, and Unicode . how encoding scheme defines the mapping between characters Common encoding schemes include UTF-8, UTF-16, ASCII, and Unicode Choosing the correct encoding is important because different encoding schemes support different sets of characters. For example, ASCII only supports the basic English alphabet (with a few more characters), while UTF-8 and UTF-16 are capable of representing characters from multiple languages and scripts. More on this later though since I know you’re eager for some code! Choosing an Encoding for String to Byte Array Conversion When converting strings to byte arrays in C#, one of the most important things to consider is character encoding. Character encoding determines the mapping between characters and byte values, and if you’re considering transforming data one way then you may want to put some thought into how to transform it back! What’s that supposed to mean? Well, if we use a data transform — let’s say taking a string and transforming it to bytes with an ASCII encoding — if there is not a mapping of a particular character to a byte representation we lose that data in the result. Now what happens if you want to go the other way and get your byte array back to a string? Data is missing! In C#, there are various encodings available, including ASCII, UTF-8, and UTF-16, each with its specific characteristics and usage scenarios. Let’s explore these encodings and see how they can be used for string to byte array conversion. ASCII Encoding ASCII encoding represents characters using 7 bits, allowing for a total of 128 different characters. It’s primarily suitable for handling basic English characters and is more space-efficient compared to other encodings. Here’s an example of converting a string to a byte array using ASCII encoding: string text = "Hello, World!";
byte[] asciiBytes = Encoding.ASCII.GetBytes(text); string text = "Hello, World!";
byte[] asciiBytes = Encoding.ASCII.GetBytes(text); In many modern applications, ASCII may not be what you’re after. This is especially true if you have users across the globe in locales that are not English. That’s not to say that ASCII *cannot* be used, but you’ll want to be careful about what data you ASCII encode so as to not lose information during the encoding transform. So keep in mind that the ASCII character set represents characters using a 7-bit encoding scheme, allowing for a total of 128 unique characters. However, with the growing need for internationalization and multilingual support, ASCII alone is insufficient to represent all characters effectively. Good thing we have some other options coming right up! UTF-8 Encoding UTF-8 encoding is a variable-length encoding scheme that can represent any Unicode character. It’s widely used for encoding text in various languages and is backward compatible with ASCII. It uses fewer bytes for representing ASCII characters but may require more bytes for non-ASCII characters. Here’s an example of converting a string to a byte array using UTF-8 encoding: string text = "Привет, мир!";
byte[] utf8Bytes = Encoding.UTF8.GetBytes(text); string text = "Привет, мир!";
byte[] utf8Bytes = Encoding.UTF8.GetBytes(text); UTF-16 Encoding UTF-16 encoding represents characters using either 2 or 4 bytes, making it capable of representing any Unicode character. It’s commonly used by applications that need to handle multilingual text or when interoperability with other systems is required. The encoding can be little-endian or big-endian, with the former being more prevalent. Here’s an example of converting a string to a byte array using UTF-16 encoding: string text = "こんにちは、世界！";
byte[] utf16Bytes = Encoding.Unicode.GetBytes(text); string text = "こんにちは、世界！";
byte[] utf16Bytes = Encoding.Unicode.GetBytes(text); UTF-8 Encoding vs UTF-16 Encoding – What’s The Difference? Each of these encodings is variable width, and their size gives them different characteristics when we consider different alphabets. Starting with UTF-8, it’s variable-width and backward-compatible with ASCII. In this encoding: ASCII characters, which are on the range U+0000 to U+007F, take only 1 byte. In the next range, code points U+0080 to U+07FF take twice as much space at 2 bytes each. Code points U+0800 to U+FFFF take one more byte, bringing us up to 3 bytes And finally, code points U+10000 to U+10FFFF take 4 bytes. ASCII characters, which are on the range U+0000 to U+007F, take only 1 byte. In the next range, code points U+0080 to U+07FF take twice as much space at 2 bytes each. Code points U+0800 to U+FFFF take one more byte, bringing us up to 3 bytes And finally, code points U+10000 to U+10FFFF take 4 bytes. This can be very effective for English text because those characters will take up minimal space. However, when it comes to Asian text it’s not ideal for the exact opposite reason. Because UTF-16 has code points from U+0000 to U+FFFF take 2 bytes and code points U+10000 to U+10FFFF are double that at 4 bytes, it’s not-so-great for English. It *does* happen to be better suited for Asian characters though. There’s even UTF-32 encoding! This is a fixed-width encoding where all of the code points take four bytes — unlike the others I’ve mentioned which are dynamic in size. This can use much more storage than the other encodings, but because of its simplicity can be much faster to operate on. Give some consideration to the alphabets you’ll need to support primarily! Best Practices for String to Byte Array Conversion When converting strings to byte arrays in C#, it’s important to follow best practices to ensure efficiency and reliability. In this section, I’ll discuss some key best practices that you should keep in mind when performing string-to-byte array conversions. Error Handling and Validation When working with encodings, it’s important to handle potential errors and validate your data to prevent unexpected behavior in your code. Ideally, you structure the flow of your application such that you know what kind of data you’re dealing with. If you can write code to avoid errors in the first place, this is preferred! We don’t want to have to rely on this but sometimes it’s outside of our control — handling errors is by using try-catch blocks. By encapsulating the conversion code within a try block, you can catch any exceptions that may occur during the conversion process and handle them gracefully. If you don’t have control over the source of the input data, this is something you might need to do for safety. Encoding Selection C# provides several encoding options for converting strings to byte arrays, such as UTF-8, UTF-16, ASCII, and more. It’s important to select the appropriate encoding based on the specific requirements of your application. Consider factors such as character sets, compatibility with other systems, and performance implications when choosing the encoding. // Example of encoding selection
string inputString = "Hello, World!";
byte[] encodedBytes = Encoding.UTF8.GetBytes(inputString); // Example of encoding selection
string inputString = "Hello, World!";
byte[] encodedBytes = Encoding.UTF8.GetBytes(inputString); Just like we saw in the previous examples, after the Encoding class we get to pick the static property holding the encoding instance. If we need to select one to pass around as a variable and parameter, you can absolutely store it in a dedicated encoding reference: Encoding selectedEncoding = Encoding.UTF8;
SomeMethod("Hello World!", selectedEncoding); Encoding selectedEncoding = Encoding.UTF8;
SomeMethod("Hello World!", selectedEncoding); Selecting the wrong encoding can have big consequences for your application! This is especially true if you save data with an encoding that will lose data resolution and you can’t reverse it… so put some care into this! Now You Know How to Convert a String to Byte Array in C#! You’re a pro now with encoding and decoding! Well, maybe not a full-on expert… but you have the basics put in front of you and some guidelines to work with. That’s a pretty good start. Remember to select the right encoding for the situation you’re dealing with. Keep in mind that you can potentially lose data resolution when using the wrong encoding, and as a result, lose data forever! If you found this useful and you’re looking for more learning opportunities, consider subscribing to my free weekly software engineering newsletter and check out my free videos on YouTube ! subscribing to my free weekly software engineering newsletter free videos on YouTube

How to Convert a String to Byte Array in C#

Decrypting the Future: Programmable Cryptography And Its Role in Modern Tech

ভবিষ্যতের ডিক্রিপ্টিং: প্রোগ্রামেবল ক্রিপ্টোগ্রাফি এবং আধুনিক প্রযুক্তিতে এর ভূমিকা

Die Zukunft entschlüsseln: Programmierbare Kryptografie und ihre Rolle in der modernen Technologie

Descifrando el futuro: criptografía programable y su papel en la tecnología moderna

Décrypter l'avenir : la cryptographie programmable et son rôle dans la technologie moderne

भविष्य को डिक्रिप्ट करना: प्रोग्रामेबल क्रिप्टोग्राफी और आधुनिक तकनीक में इसकी भूमिका

未来を解読する: プログラム可能な暗号化と現代技術におけるその役割

미래의 암호 해독: 프로그래밍 가능한 암호화와 현대 기술에서의 역할

Descriptografando o futuro: criptografia programável e seu papel na tecnologia moderna

Расшифровка будущего: программируемая криптография и ее роль в современных технологиях

Geleceğin Şifresini Çözmek: Programlanabilir Kriptografi ve Modern Teknolojideki Rolü

Giải mã tương lai: Mật mã có thể lập trình và vai trò của nó trong công nghệ hiện đại

解密未来：可编程加密技术及其在现代技术中的作用

Narrative Resilience in Online Conflict Coverage: Insights from Twitter Analysis

Cryptography and the Golden Security Line

Mastering the Craft of Transforming Data into Engaging Visual Narratives in User Interfaces

The world of work is changing––rapidly. Thanks to the advent and rapid adoption of generative AI, tech stocks had a bumper 2023, with the share price of the Big Five ( , , , , and ) leaping by 60% from the start of that year. 2023 was also a , which saw its shares triple in price. Alphabet Amazon Apple Meta Microsoft smash success for chipmaker, Nvidia And while adoption isn’t even across the board yet, corporate America is sitting up and taking notice. At JPMorgan Chase, for example, the company said it had 600 machine-learning engineers, with AI working on more than 300 different internal applications. A recent found that 71% of C-suite leaders said that advancing their use of AI is either a high or medium priority for their organization, with 62% of executives noting that they are worried they’re not moving fast enough with AI. study from UKG That has led to some companies doing “AI washing”, aka exaggerating the amount of AI technology they use in their products or services. While some companies might not be entirely truthful about their motivations and use of artificial intelligence or machine learning technologies, what is increasingly true is that in an AI future, tech workers will need to upskill. The World Economic Forum (WEF) says that “the 2020s are becoming the decade of development for employees and employers in terms of skills and career progression”. Its most recent identified that 44% of workers’ skills will be disrupted in the next five years, and six in 10 staff members will require training before 2027. Future of Jobs Report Upskilling for free Key skills to acquire will include new technology capabilities, but the WEF has also identified key human skills such as analytical and creative thinking, as well as resilience, flexibility and agility, motivation, and self-awareness. AKA, the things that artificial intelligence (so far) can’t do. If you’re thinking about some upskilling, then one great place to begin is (OCW) platform. MIT’s OpenCourseWare A free and open space to learn, OCW contains material from thousands of MIT courses across every MIT department and degree program, ranging from the introductory to the most advanced graduate level. Courses include a syllabus, instructional material (such as lecture notes or a reading list), and learning activities like assignments or exams. You can use all the courses at your own pace, and there is no enrollment fee or date to complete by––useful for those working in busy, pressurized roles. A huge range of computer science, engineering, and technology courses are available, across topics such as machine learning, generative AI, Python, deep learning, cognitive robotics, data science, and much more, all for free. If upskilling to get a new job is on your 2024 agenda, then the courses above are a great start. And your job hunt is made far easier too thanks to the . It contains thousands of interesting tech roles, like the three below. HackerNoon Job Board DataAnnotation is seeking a to train its AI chatbots to code. This flexible, part-time or full-time role allows you to choose the projects you want to work on, involves coming up with diverse problems and solutions for a coding chatbot, writing high-quality answers and code snippets, and evaluating the code quality produced by the AI model. You will need to be proficient in at least one programming language and able to solve coding problems (LeetCode, HackerRank). Software Developer Adobe is looking for a in San Jose. The salary range for this position is $135,200 - $250,900, and you will have either research or industry experience in building scalable data and ML platforms. In this role, you will have the opportunity to build a new GenAI foundation to change the creative landscape and will collaborate with data platform engineers and architects to seamlessly integrate low latency data pipelines into the ML platform for model training, as well as writing high quality, product level code that is easy to maintain and test following standard methodologies. Other tasks include designing and implementing reusable and scalable data-loading framework to support video and audio foundational model training in large-scale and distributed environments. Machine Learning Engineer Python programmers looking for a new role can consider this role at SAIC in Chantilly. You’ll be experienced, results-oriented, and mission-driven, with a specialized focus on data engineering in order to perform data model design, data formatting, and ETL development optimized for efficient storage, access, and computation in support of national security objectives. A solid understanding and experience with developing, automating, and enhancing all parts of the data pipeline to include ingestion, processing, storage, and exposing data for consumption is required, along with a focus on improving inefficient tooling and adopting new transformative technologies while maintaining operational continuity. Python Developer Want a great new tech role? Discover it now on the HackerNoon Job Board right By Kirstie McDermott

MIT Has Released a Bunch of Free Courses: Discover 5 To Boost Your Tech Career

এমআইটি একগুচ্ছ বিনামূল্যের কোর্স প্রকাশ করেছে: আপনার কারিগরি ক্যারিয়ারকে উন্নত করতে 5 আবিষ্কার করুন

Das MIT hat eine Reihe kostenloser Kurse veröffentlicht: Entdecken Sie 5, um Ihre Tech-Karriere voranzutreiben

El MIT ha lanzado una serie de cursos gratuitos: descubra 5 para impulsar su carrera tecnológica

Le MIT a publié un certain nombre de cours gratuits : découvrez-en 5 pour booster votre carrière technologique

एमआईटी ने कई मुफ्त पाठ्यक्रम जारी किए हैं: अपने तकनीकी करियर को बढ़ावा देने के लिए 5 पाठ्यक्रमों के बारे में जानें

MIT が無料コースを多数公開: テクノロジーのキャリアを伸ばす 5 つのコースをご紹介します

MIT가 다양한 무료 강좌를 출시했습니다: 기술 경력을 향상하는 5가지를 찾아보세요

O MIT lançou vários cursos gratuitos: descubra 5 para impulsionar sua carreira tecnológica

MIT выпустил несколько бесплатных курсов: откройте для себя 5, которые помогут улучшить вашу техническую карьеру

MIT Bir Çok Ücretsiz Kurs Yayımladı: Teknoloji Kariyerinizi Yükseltecek 5 Kursu Keşfedin

MIT đã phát hành một loạt khóa học miễn phí: Khám phá 5 khóa học để thúc đẩy sự nghiệp công nghệ của bạn

麻省理工学院发布了一系列免费课程：探索 5 门课程，助你提升科技职业生涯

This paper is available on arxiv under CC 4.0 license.
Authors:
(1) Sebastien Ferr ´e, Univ Rennes, CNRS, Inria, IRISA Campus de Beaulieu, 35042 Rennes, France and ferre@irisa.fr. Table of Links Abstract & Introduction Abstraction and Reasoning Corpus (ARC) Related Work Object-centric Models for ARC Grids MDL-based Model Learning Evaluation Conclusion and Perspectives, References & Supplementary Materials 3 Related Work The ARC benchmark is recent and not many approaches have been published so far. All those we know define a DSL (Domain-Specific Language) of programs that transform an input grid into an output grid, and search for a program that is correct on the training examples [10,3,23,2]. The differences mostly lie in the primitive transformations (prior knowledge) and in the search strategy. It is tempting to define more and more primitives like the Kaggle winner did, hence more prior knowledge, but this means a less intelligent system according to Chollet’s measure. To guide the search in the huge program space, those approaches use either grammatical evolution [10], neural networks [3], search tree pruning with hashing and Tabu list [23], or stochastic search trained on solved tasks [2]. A difficulty is that the output grids are generally only used to score a candidate program so that the search is kind of blind. Alford [3] improves this with a neural-guided bi-directional search that grows the program in both directions, from input and output. Xu [23] compares the in-progress generated grid to the expected grid but this limits the approach to the tasks whose output grids have the same size and same objects as the input grids. DSL-based approaches have a scaling issue because the search space increases exponentially with the number of primitives. Ainooson [2] alleviates this difficulty by defining high-level primitives that embody specialized search strategies. We compare and discuss their performance in the evaluation section. Johnson et al. [14] report on a psychological study of ARC. It reveals that humans use object-centric mental representations to solve ARC tasks. This is in contrast with existing solutions that are based on grid transformations. Interestingly, the tasks that are found the most difficult by humans are those based on logics (e.g., an exclusive-or between grids) and symmetries (e.g., rotation), precisely those most easily solved by DSL-based approaches. The study exhibits two challenges: (1) the need for a large set of primitives, especially about geometry; (2) the difficulty to identify objects, which can be only visible in part due to overlap or occlusion. A valuable resource is LARC, for Language-annotated ARC [1], collected by crowd-sourcing. It provides for most training tasks one or several natural programs that confirm the object-centric and declarative nature of human representations. A natural program is short textual descriptions produced by a participant that could be used by another participant to generate test output grids (without access to the training examples). Beyond the ARC benchmark, a number of work has been done in the domain of program synthesis, which is also known as program induction or programming by examples (PbE) [17]. An early approach is Inductive Logic Programming (ILP) [19], where target predicates are learned from symbolic representations. PbE is used in the FlashFill feature of Microsoft Excel 2013 to learn complex string processing formulas from a few examples [18]. Dreamcoder [8] alternates a wake phase that uses a neurally guided search to solve tasks, and a sleep phase that extends a library of abstractions to compress programs found during wake. Bayesian program learning was shown to outperform deep learning at parsing and generating handwritten world’s alphabets [16]. This paper is available on Arxiv under CC 4.0 license. This paper is available on arxiv under CC 4.0 license. Authors: (1) Sebastien Ferr ´e, Univ Rennes, CNRS, Inria, IRISA Campus de Beaulieu, 35042 Rennes, France and ferre@irisa.fr. This paper is available on arxiv under CC 4.0 license. Authors: Authors: (1) Sebastien Ferr ´e, Univ Rennes, CNRS, Inria, IRISA Campus de Beaulieu, 35042 Rennes, France and ferre@irisa.fr. Table of Links Abstract & Introduction Abstract & Introduction Abstraction and Reasoning Corpus (ARC) Abstraction and Reasoning Corpus (ARC) Related Work Related Work Object-centric Models for ARC Grids Object-centric Models for ARC Grids MDL-based Model Learning MDL-based Model Learning Evaluation Evaluation Conclusion and Perspectives, References & Supplementary Materials Conclusion and Perspectives, References & Supplementary Materials 3 Related Work The ARC benchmark is recent and not many approaches have been published so far. All those we know define a DSL (Domain-Specific Language) of programs that transform an input grid into an output grid, and search for a program that is correct on the training examples [10,3,23,2]. The differences mostly lie in the primitive transformations (prior knowledge) and in the search strategy. It is tempting to define more and more primitives like the Kaggle winner did, hence more prior knowledge, but this means a less intelligent system according to Chollet’s measure. To guide the search in the huge program space, those approaches use either grammatical evolution [10], neural networks [3], search tree pruning with hashing and Tabu list [23], or stochastic search trained on solved tasks [2]. A difficulty is that the output grids are generally only used to score a candidate program so that the search is kind of blind. Alford [3] improves this with a neural-guided bi-directional search that grows the program in both directions, from input and output. Xu [23] compares the in-progress generated grid to the expected grid but this limits the approach to the tasks whose output grids have the same size and same objects as the input grids. DSL-based approaches have a scaling issue because the search space increases exponentially with the number of primitives. Ainooson [2] alleviates this difficulty by defining high-level primitives that embody specialized search strategies. We compare and discuss their performance in the evaluation section. Johnson et al. [14] report on a psychological study of ARC. It reveals that humans use object-centric mental representations to solve ARC tasks. This is in contrast with existing solutions that are based on grid transformations. Interestingly, the tasks that are found the most difficult by humans are those based on logics (e.g., an exclusive-or between grids) and symmetries (e.g., rotation), precisely those most easily solved by DSL-based approaches. The study exhibits two challenges: (1) the need for a large set of primitives, especially about geometry; (2) the difficulty to identify objects, which can be only visible in part due to overlap or occlusion. A valuable resource is LARC, for Language-annotated ARC [1], collected by crowd-sourcing. It provides for most training tasks one or several natural programs that confirm the object-centric and declarative nature of human representations. A natural program is short textual descriptions produced by a participant that could be used by another participant to generate test output grids (without access to the training examples). Beyond the ARC benchmark, a number of work has been done in the domain of program synthesis, which is also known as program induction or programming by examples (PbE) [17]. An early approach is Inductive Logic Programming (ILP) [19], where target predicates are learned from symbolic representations. PbE is used in the FlashFill feature of Microsoft Excel 2013 to learn complex string processing formulas from a few examples [18]. Dreamcoder [8] alternates a wake phase that uses a neurally guided search to solve tasks, and a sleep phase that extends a library of abstractions to compress programs found during wake. Bayesian program learning was shown to outperform deep learning at parsing and generating handwritten world’s alphabets [16]. This paper is available on Arxiv under CC 4.0 license. Arxiv

Stories, News & Mentions about ALPHABET

Alphabet’s Stories on HackerNoon

Alphabet's News & Mentions