J2 Health

j2health

There has been a lot of buzz about OpenAI GPT-3, now having the largest neural network. Does it mean the AI problem has been solved? Yes, it has a large dataset, but we still don’t know how it learns. OpenAI Basics OpenAI Inc is a non-profit arm of Open.AI LP whose goal is to create a ‘friendly AI’ that will benefit humanity. Open.AI has several different offerings: - an AI system that can create realistic images and art from a description in natural language DALL•E 2 - Generative Pre-trained Transformer is a language model that leverages deep learning to generate human-like text GPT-3 - an updated model that produces less offensive language and fewer mistakes overall but may also generate misinformation InstructGPT - Contrastive Language-Image Pre-training. It recognizes visual concepts in images and associates them with their names. CLIP How are the Models Trained? OpenAI GPT-3 is trained on 500 billion words by using the following datasets: The dataset contains data collected from over 8 years of web crawling Common Crawl is the text of webpages from all outbound Reddit links of posts with 3+ upvotes WebText2 are two internet-based books corpora Books 1 & Books2 pages in the English language Wikipedia Dataset breakdown and training distribution Dataset Tokens Weight in Training Common Crawl 410 billion 60% WebText2 19 billion 22% Books1 12 billion 8% Books2 55 billion 8% Wikipedia 3 billion 3% Training Models can be done using the following methods: This is where we give between 10-100 contexts to a model and expect the model to determine what comes next. Few-shot (FS). This is quite similar to FS. However, an example is given without any training. Context is given to the model to determine what word comes next. One-shot (1S). Zero-Shot (0S) The model predicts the answer given. The idea is that during training, the model has seen enough samples to determine what word comes next. Only the last context is allowed, making this setting difficult. Bias is inevitable Training the model involves taking large bodies of text for GPT-3 and images for DALL•E from the internet. This is where the problem occurs. The model encounters the best and worst. To counter this, OpenAI created InstructGPT, While training InstructGPT, Open.ai hired 40 people to rate the responses and would reward the model accordingly. DALL•E 2 Open.ai outlines the they currently encounter: Risks and Limitations “Use of DALL·E 2 has the potential to harm individuals and groups by reinforcing stereotypes, erasing or denigrating them, providing them with disparately low quality performance, or by subjecting them to indignity.’’ This is what DALL•E 2 believes a ‘CEO’ looks like: This is what DALL•E 2 believes a ‘flight attendant’ looks like: To reduce bias, OpenAI has recruited external experts to provide feedback. GPT-3 Gender Bias To test bias, I borrowed a list of from . You can use the OpenAI to test it for yourself. The results prove to be quite Gender bias prompts Jenny Nicholson playground interesting. Phrases: female/male employee women/men in the c-suite any woman/man knows women/men entering the workforce should know Religious Bias Gender and Race are biases that have been studied in the past. However, a recent reveals that GPT-3 also has religious bias. The following was found: paper Muslim mapped to “terrorist” in 23% of test cases Jewish mapped to “money” in 5% of test cases CLIP Race, Gender, and Age Bias CLIP performs well on classification tasks, as you have already seen in this article. It uses as its dataset to train the model. This is due to the images it is scraping from the internet. However, the model breaks down when it classifies age, gender, race, weight, and so on. This means the AI tools used to generate new art can continue perpetuating recurring  stereotypes. ImageNet OpenAI can be used to improve content generation. But as long as the datasets are being trained by scraping existing internet, we will build biases against age, gender, race, and more into technology. We must take precautions when using the internet. The information that goes into the AI must be filtered, or harmful stereotypes will never be erased.

Here’s How OpenAI is Perpetuating Unhealthy Stereotypes

Así es como OpenAI está perpetuando los estereotipos poco saludables

Voici comment OpenAI perpétue des stéréotypes malsains

यहां बताया गया है कि कैसे OpenAI अस्वास्थ्यकर रूढ़ियों को कायम रख रहा है

OpenAI が不健康な固定観念をどのように永続させているか

Veja como o OpenAI está perpetuando estereótipos prejudiciais

Đây là cách OpenAI đang tồn tại những định kiến không lành mạnh

以下是 OpenAI 如何延续不健康的刻板印象

Over the past few months, I have increasingly noticed that more and more developer teams that I help coach are turning their backs on monorepos, instead choosing to liberate code to separate repositories. For me, this couldn’t have come sooner. Don’t get me wrong, I have seen examples of monorepos work well in the past - a few years ago I worked on a scope-limited system written in which was deployed in three different environments with significant amounts of code reuse and a single team owning the entire codebase. In circumstances like this, monorepos made sense for this project. Go Unfortunately, over the past few years, companies have blindly adopted monorepos as an architectural pattern for all organisational code without considering the individual circumstances of what they’re building. In this blog post, I wanted to present some of the reasons why I think this is a firmly unhealthy engineering practice. Appealing to Big Tech Authority One of the most common arguments used by those who advocate monorepos is an appeal to the authority of big tech companies who have adopted them. If Google, Facebook, etc use monorepos, surely shouldn’t we too? This argument ignores the unique resources and challenges that these companies face. Quite simply, a company with 100s of engineers (or fewer) isn’t likely to have the developer productivity resources needed to allow monorepos to scale. For example, Google internally has that helps manage the review and merge process of cross-team Pull Requests in a monorepo. With tools like Rosie not being in the public domain, many engineering teams have learned the hard way that they don’t have the resources to support such architectural patterns. According to the tweets of the former VP of Architecture at Monzo (a UK online banking start-up), . a tool called Rosie they experienced this very problem Instances like this provide a stark reminder that the fact a big technology company does something is not a substitute for your own professional engineering judgment. Monorepos Create Tightly Coupled Code Google’s consists of seven years of research and data from more than 32,000 professionals across the industry, in companies of different sizes. Their research reports in and were clear; loosely-coupled software architecture is one of the strongest predictors of successful continuous delivery. DORA State of DevOps Report research 2021 2019 Monorepos add significant complexity to the deployment pipeline of software. As the monorepo scales, deploying all software for every change quickly becomes impractical. Slow builds or flaky tests in one service can grind the entire deployment pipeline to a halt. Therefore teams will inevitably face the challenge of adding infrastructure to work out what service needs to be rebuilt and deployed, and which can stay the same. This adds complexity, risk, and overhead to the process of deploying changes to software. As the DORA research describes, for effective high-trust engineering organisations, "teams can scale, fail, test, and deploy independently of one another". This is dependent on loosely-coupled services being separated by strictly-defined interfaces. Removing the source code boundaries between different services removes the strictly-defined interfaces necessary for this. No Architecture Is Better than Bad Architecture In my own experience, the outcomes of architectural practices like monorepos are worse than if the organisation had no centrally proscribed architectural mandate and simply allowed the systems architecture to emerge organically. This raises the question, why is this done in the first place? Conway’s Law is an adage that says: “Any organization that designs a system (defined broadly) will produce a design whose structure is a copy of the organization's communication structure.” In low-trust teams, a centralised architecture unit or DevOps team might emerge to mandate practices instead of trusting engineering teams to use their own judgments. It is therefore important to look beyond just technical factors and look to the organisational reasons that led to unhealthy software development practices. The character of in films reminds me much of how unaccountable software architects end up conducting themselves. Without being challenged on whether their system met the outcomes of what it was built for, they merely end up trying to build grand structures for the sake of it. The Architect The Matrix By contrast, the engineering teams actually building the functionality at hand are constantly held accountable for the work they produce. Giving individual engineering managers (and thereby, software developers) the freedom and responsibility for the technical architecture of their services, you achieve better outcomes.

Thinking Critically About Monorepos and Why Its an Unhealthy Engineering Practice

: Vested Interest Disclosure The author's an independent contributor, and while HackerNoon has reviewed the story for quality, the claims hereon belong to the author. #DYOR Hi Ethan, welcome to our series “Behind the Startup.” Please tell us about yourself and the story behind WonderHero? Ishan Pandey: Hi Ishan, thank you for having me. It’s an honor to be part of this series! I’m Ethan, Co-Founder, and CEO of WonderHero. WonderHero was founded from a spark of luck. When I was a teenager, I was a game addict and became really good at games in a short amount of time. I was told to stop due to my mum nagging at me, saying, “You better quit games or risk being a loser for the rest of your life”. Ethan Ng: I believe a lot of folks can relate to this. So when my co-founder approached me for this project to lead as the CEO, after playing the alpha game for 3.5 hours, I knew this was it. Since then, I have set forth the mission to fulfill the dream that I couldn’t fulfill as a teenager for many others out there, only this time, they have an opportunity to play and earn. Please tell us a little bit about WonderHero and the P2E NFT based gaming model. Ishan Pandey: WonderHero is a Play-And-Earn RPG for iOS & Android, where players control a team of 4 heroes in tactical turn-based battles to win NFTs and Crypto Tokens. They can trade the NFTs (Hero Skins, Weapons, and Equipment) on our in-built Marketplace and liquidate the Tokens on exchanges for income. Ethan Ng: Besides trading NFTs on the marketplace, players can also stake NFT to earn more Token rewards. To start playing WonderHero, players need at least a minimum of 3 playable Heroes, or they can join a Guild as Scholar to play for free. Can you explain to our readers how crypto gaming guilds differ from traditional gaming guilds? Ishan Pandey: That is a good one. In my opinion, fundamentally, the key difference between crypto gaming guilds and traditional ones is the players’ intent. Ethan Ng: Traditional gaming guilds are formed out of pure passion and interest in a particular game, where players thrive by partying up to progress faster in-game. Gamers volunteered to play a role in their guilds or are promoted through their contributions. With crypto gaming guilds, on the other hand, the primary intent of gaming started as earning income, which introduces a different gaming behavior, mindset, and structure. This means more players are playing for the earnings than for the game itself. The second difference is that while both types of gaming guilds are supportive pillars for the gaming ecosystem to thrive, some crypto gaming guilds have become massive with investments from VCs and other entities. Besides being able to channel players to a P2E game, guilds can also make a financial impact by investing in new projects, purchasing NFT assets (like land) in the game. How can we successfully build a sustainable gaming economy while the industry is undergoing a pragmatic shift in light of the budding popularity of the NFT gaming model? Ishan Pandey: It is truly exciting to be in this space, as we are so early on and everyone is figuring out what is the best way to build a sustainable gaming economy. Much like the early days when games were on consoles and the industry was finding the best model to bring games onto social media networks and the free-to-play model for mobile. Eventually, the industry figured out how to build sustainable free-to-play models, and there were best practices to model after. Ethan Ng: The P2E space is undergoing such a transition too, and every project is experimentation and trying to figure out a sustainable model, but I believe player retention is at the core of a sustainable gaming economy. The game needs to be fun, enjoyed by its community, and keep its players coming back for more. The majority of current P2E gamers are drawn in on the earning aspect, which is unhealthy in the long run for games because players and guilds could move from one game to another based on the earnings potential. The way sustainability works would mean that P2E games need to be able to attract gamers and grow the community even without the earning aspects of it. The earning aspect is more of a means to amplify the growth. OpenSea, currently the largest global NFT marketplace, was hacked, and consumers lost their NFTs. In this regard, what regulatory measures, you suggest, should be incorporated in terms of regulating NFTs to avoid such cyberattacks in the future? Ishan Pandey: I think there are two parts to this: Ethan Ng: Firstly, the majority of the consumers who are victims of hacks are relatively new to the crypto space and are unaware of the possible phishing hack scams. This means information sharing and education for consumers to learn how to protect themselves is much needed as part of their first lessons, much like how we were taught not to tell someone else passwords or sensitive personal information. I believe this is something all projects could do as a collective effort to make this space safer for newcomers. Secondly, vulnerability testing will be an essential practice for projects to keep a lookout for potential security threats against their community and ecosystem. What does the roadmap ahead look like for WonderHero and the NFT-based gaming ecosystem as a whole? Ishan Pandey: For the NFT-based gaming ecosystem, I believe we are entering a consolidation phase. There is a huge influx of projects in the past few months due to the hype and now we are in the midst of a bear market, which means at the end of this phase, we will see which are the projects that didn’t make it through and which are the ones who survived and is thriving with their economy. Ethan Ng: There are also many exciting new projects with promising potential, exploring various mechanics and economic models, which will be interesting to watch how it will turn out. I’m confident that NFT-based games will enter a new phase with better gameplay and a stronger model that revolves around game retention and token utility at the end of this bear market period. As for WonderHero, our team is refining our economic model, increasing NFT and token utilities. The guilds have received the guild management system well and we are continuously working to make it better. At the same time, besides having new content for the game, we are also exploring ways to increase the competitiveness of our PVP as well as other game modes. There are lots to do and there are also many great suggestions from our community that we are evaluating before deciding to add to our roadmap.

"Current P2E Gamers are drawn in on the earning aspect, which is unhealthy" - Ethan Ng

I think OpenAI is not being honest about the diminishing returns of scaling AI with data and compute alone. I think they are also putting a lot of the economy, the world and this entire industry in jeopardy by not talking more openly about the topic. At first, I believed what they told us, that all you need to do is add more compute power and more data, and LLMs as well as other models will simply get better. That this relationship between the models, their compute and data could grow linearly until the end of time. The leap from GPT-3 and GPT-3.5 was immense. The leap from GPT-3.5 to GPT-4 seemed like clear evidence that this presumption was correct. But then things got weird. Instead of releasing a model called GPT-5 or even GPT-4.5, they released GPT-4-turbo. GPT-4-turbo is not as intelligent as GPT-4 but it is much faster and it's cheaper. That all makes sense. But then, this trend kept going. After GPT-4-turbo, OpenAI's next release was GPT-4o. GPt-4o is more or less just as intelligent as GPT-4-turbo, but it is even faster and even cheaper. The functionality that really sold us, however, was its ability to talk and understand things via audio and its speed. However, take note, at this point in our story, GPT-4-turbo is not more intelligent than GPT-4 and GPT-4o is not more intelligent than GPT-4-turbo. And none of them are more intelligent than GPT-4. Their next and most recent release was GPT-o1. GPT-o1 can perform better than GPT-4 onsome tasks. But that's because o1 is not really a single model. GPT-o1 is actually a black box of multiple lightweight LLM models working together. Perhaps o1 is even better described as software or middleware than it is an actual model. You give it a question, it comes up with an answer, and then it repeatedly uses other models tasked with checking the answer to make sure it’s right, and it disguises all of these operations. It does all of this very, very quickly. Why not just make an LLM that's more powerful than GPT-4? Why resort to such cloak-and-dagger techniques to achieve new releases? GPT-4 came out 2 years ago, we should be well beyond its capabilities by now. Well, Noam Brown, a researcher at OpenAI had something to say on why they went this route with o1 at TED AI. He said “It turned out that having a bot think for just 20 seconds in a hand of poker got the same boosting performance as scaling up the model by 100,000x and training it for 100,000 times longer,” Now stop and really think about what is being said there. A bot thinking for 20 seconds is as good as a bot trained 100,000 times longer with 100,000 times more computing power.  If the scaling laws are infinite, that math is impossible. Something is either wrong here or someone is lying. Why does all of this matter? OpenAI is worth 150 billion dollars and the majority of that market cap is based on projections that depend on the improvement of models over time. If AI is only as good as it is today, that’s still an interesting future, but that’s not what’s being sold to investors by AI companies whose entire IP is their model. That also changes the product roadmap of many other companies that depend on the continued advancement of their LLMs to build their own products. OpenAI’s goal and ambitions of AGI are severely delayed if this is all true. A Hypothesis The reason LLMs are so amazing is because of a higher level philosophical phenomenon that we never considered, that language inherently possesses an extremely large amount of context and data about the world within even small sections of text. Unlike pixels in a picture or video, words in a sentence implicitly describe one another. A completely cohesive sentence is by definition, “rational”. Whether or not it’s true is a very different story and a problem that transcends language alone. No matter how much text you consume, “truth” and “falsehoods” are not simply linguistic concepts. You can say something is completely rational but in no way “true”. It is at this point LLMs will consistently hit a brick wall. Over the last 12 months, I’d like to formally speculate that behind closed doors there have been no huge leaps in LLMs at OpenAI, GrokAI, or at Google. To be specific I don’t think anyone, anywhere has made any LLM that is even 1.5X better than GPT-4. At OpenAI it seems that high-level staff are quitting. Right now they’re saying it’s because of safety but I’m going to put my tinfoil hat on now and throw an idea out there. They are aware of this issue and they’re jumping ship before it’s too late. Confirmation I started discussing this concern with friends 3 months ago. I was called many names haha. But in the last 3 weeks, a lot of the press has begun to smell something fishy: OpenAI is no longer releasing Orion (GPT-5) because it did not meet expected performance benchmarks and it is seeing diminishing returns. (https://www.theinformation.com/articles/openai-shifts-strategy-as-rate-of-gpt-ai-improvements-slows)


Bloomberg reports that OpenAI, Google and Anthropic are all having struggles making more advanced AI. (https://www.bloomberg.com/news/articles/2024-11-13/openai-google-and-anthropic-are-struggling-to-build-more-advanced-ai) What can we do about it? It’s hard to recommend a single solution. The tech behind o1 is proof that even low-performance models can be repurposed to do complicated operations. But that is not a solution to the problem of AI scaling. I think there needs to be substantial investment and rapid testing of new model architectures. We also have run out of data and need new ways of extrapolating usable data for LLMs to be trained on. Perhaps using multidimensional labeling that helps guide its references for truthful information directly. Another good idea could be to simply continue fine-tuning LLMs for specific use cases like math, science, and healthcare running and using AI agent workflows, similar to o1. It might give a lot of companies wiggle room until a new architecture arises. This problem is really bad but I think that the creativity in machine learning and software development it will inspire will be immense. Once we get over this hurdle, we’ll certainly be well on schedule for AGI and perhaps ASI. I think OpenAI is not being honest about the diminishing returns of scaling AI with data and compute alone. I think they are also putting a lot of the economy, the world and this entire industry in jeopardy by not talking more openly about the topic. At first, I believed what they told us, that all you need to do is add more compute power and more data, and LLMs as well as other models will simply get better. That this relationship between the models, their compute and data could grow linearly until the end of time. The leap from GPT-3 and GPT-3.5 was immense. The leap from GPT-3.5 to GPT-4 seemed like clear evidence that this presumption was correct. But then things got weird. Instead of releasing a model called GPT-5 or even GPT-4.5, they released GPT-4-turbo. GPT-4-turbo is not as intelligent as GPT-4 but it is much faster and it's cheaper. That all makes sense. But then, this trend kept going. After GPT-4-turbo, OpenAI's next release was GPT-4o. GPt-4o is more or less just as intelligent as GPT-4-turbo, but it is even faster and even cheaper. The functionality that really sold us, however, was its ability to talk and understand things via audio and its speed. However, take note, at this point in our story, GPT-4-turbo is not more intelligent than GPT-4 and GPT-4o is not more intelligent than GPT-4-turbo. And none of them are more intelligent than GPT-4. Their next and most recent release was GPT-o1. GPT-o1 can perform better than GPT-4 on some tasks. But that's because o1 is not really a single model. GPT-o1 is actually a black box of multiple lightweight LLM models working together. Perhaps o1 is even better described as software or middleware than it is an actual model. You give it a question, it comes up with an answer, and then it repeatedly uses other models tasked with checking the answer to make sure it’s right, and it disguises all of these operations. It does all of this very, very quickly. some Why not just make an LLM that's more powerful than GPT-4? Why resort to such cloak-and-dagger techniques to achieve new releases? GPT-4 came out 2 years ago, we should be well beyond its capabilities by now. Well, Noam Brown, a researcher at OpenAI had something to say on why they went this route with o1 at TED AI. He said “It turned out that having a bot think for just 20 seconds in a hand of poker got the same boosting performance as scaling up the model by 100,000x and training it for 100,000 times longer,” Now stop and really think about what is being said there. A bot thinking for 20 seconds is as good as a bot trained 100,000 times longer with 100,000 times more computing power.  If the scaling laws are infinite, that math is impossible. Something is either wrong here or someone is lying. Why does all of this matter? OpenAI is worth 150 billion dollars and the majority of that market cap is based on projections that depend on the improvement of models over time. If AI is only as good as it is today, that’s still an interesting future, but that’s not what’s being sold to investors by AI companies whose entire IP is their model. That also changes the product roadmap of many other companies that depend on the continued advancement of their LLMs to build their own products. OpenAI’s goal and ambitions of AGI are severely delayed if this is all true. A Hypothesis The reason LLMs are so amazing is because of a higher level philosophical phenomenon that we never considered, that language inherently possesses an extremely large amount of context and data about the world within even small sections of text. Unlike pixels in a picture or video, words in a sentence implicitly describe one another. A completely cohesive sentence is by definition, “rational”. Whether or not it’s true is a very different story and a problem that transcends language alone. No matter how much text you consume, “truth” and “falsehoods” are not simply linguistic concepts. You can say something is completely rational but in no way “true”. It is at this point LLMs will consistently hit a brick wall. Over the last 12 months, I’d like to formally speculate that behind closed doors there have been no huge leaps in LLMs at OpenAI, GrokAI, or at Google. To be specific I don’t think anyone, anywhere has made any LLM that is even 1.5X better than GPT-4. At OpenAI it seems that high-level staff are quitting. Right now they’re saying it’s because of safety but I’m going to put my tinfoil hat on now and throw an idea out there. They are aware of this issue and they’re jumping ship before it’s too late. Confirmation Confirmation I started discussing this concern with friends 3 months ago. I was called many names haha. But in the last 3 weeks, a lot of the press has begun to smell something fishy: OpenAI is no longer releasing Orion (GPT-5) because it did not meet expected performance benchmarks and it is seeing diminishing returns. (https://www.theinformation.com/articles/openai-shifts-strategy-as-rate-of-gpt-ai-improvements-slows) Bloomberg reports that OpenAI, Google and Anthropic are all having struggles making more advanced AI. (https://www.bloomberg.com/news/articles/2024-11-13/openai-google-and-anthropic-are-struggling-to-build-more-advanced-ai) OpenAI is no longer releasing Orion (GPT-5) because it did not meet expected performance benchmarks and it is seeing diminishing returns. (https://www.theinformation.com/articles/openai-shifts-strategy-as-rate-of-gpt-ai-improvements-slows) OpenAI is no longer releasing Orion (GPT-5) because it did not meet expected performance benchmarks and it is seeing diminishing returns. ( https://www.theinformation.com/articles/openai-shifts-strategy-as-rate-of-gpt-ai-improvements-slows ) OpenAI is no longer releasing Orion (GPT-5) because it did not meet expected performance benchmarks and it is seeing diminishing returns. https://www.theinformation.com/articles/openai-shifts-strategy-as-rate-of-gpt-ai-improvements-slows Bloomberg reports that OpenAI, Google and Anthropic are all having struggles making more advanced AI. (https://www.bloomberg.com/news/articles/2024-11-13/openai-google-and-anthropic-are-struggling-to-build-more-advanced-ai) Bloomberg reports that OpenAI, Google and Anthropic are all having struggles making more advanced AI. ( https://www.bloomberg.com/news/articles/2024-11-13/openai-google-and-anthropic-are-struggling-to-build-more-advanced-ai ) Bloomberg reports that OpenAI, Google and Anthropic are all having struggles making more advanced AI. https://www.bloomberg.com/news/articles/2024-11-13/openai-google-and-anthropic-are-struggling-to-build-more-advanced-ai What can we do about it? It’s hard to recommend a single solution. The tech behind o1 is proof that even low-performance models can be repurposed to do complicated operations. But that is not a solution to the problem of AI scaling. I think there needs to be substantial investment and rapid testing of new model architectures. We also have run out of data and need new ways of extrapolating usable data for LLMs to be trained on. Perhaps using multidimensional labeling that helps guide its references for truthful information directly. Another good idea could be to simply continue fine-tuning LLMs for specific use cases like math, science, and healthcare running and using AI agent workflows, similar to o1. It might give a lot of companies wiggle room until a new architecture arises. This problem is really bad but I think that the creativity in machine learning and software development it will inspire will be immense. Once we get over this hurdle, we’ll certainly be well on schedule for AGI and perhaps ASI. software development

Is AI Progress Slowing? The Scaling Debate OpenAI Doesn’t Want to Have

¿Se está ralentizando el progreso de la IA? El debate sobre la escalabilidad que OpenAI no quiere tener

האם התקדמות הבינה המלאכותית מואטת? הדיון בקנה מידה ש-OpenAI לא רוצה שיהיה

Apakah Kemajuan AI Melambat? Perdebatan Skala yang Tidak Diinginkan OpenAI

Il progresso dell'intelligenza artificiale sta rallentando? Il dibattito sulla scalabilità che OpenAI non vuole avere

AIの進歩は鈍化しているのか？OpenAIが望まないスケーリングの議論

Bokóli ya AI ezali kokita? Débat ya Échelle OpenAI Elingi Ezala Te

Vertraagt de vooruitgang van AI? Het schaaldebat dat OpenAI niet wil hebben

Spomaľuje sa pokrok AI? Debata o škálovaní OpenAI nechce mať

Saktar AI-utvecklingen ner? Skalningsdebatten som OpenAI inte vill ha

AI ösüşi haýalmy? Giňeldilen jedel OpenAI islemeýär

Чи сповільнюється розвиток ШІ? Дебати щодо масштабування, яких OpenAI не хоче мати

Ingabe Intuthuko ye-AI Iyahamba Kancane? I-Scaling Debate OpenAI Ayifuni Ukuba Nayo

This morning, after watching President-Elect Donald Trump's announcement regarding his plan to "shatter" the censorship regime, I experienced a profound epiphany. The discussion surrounding free speech and the manipulation of information sparked thoughts about the potential for change in the current political landscape, particularly concerning figures like Edward Snowden. For years, Snowden has been a polarizing figure—celebrated by many as a whistleblower and condemned by others as a traitor. His revelations about government surveillance practices opened a critical dialogue about privacy, security, and the role of government in the lives of citizens. Yet, despite the importance of his contributions to public discourse, he has remained in exile, away from his homeland. In Trump's speech, he emphasized the need to dismantle what he described as a "censorship cartel" created by various entities, including government agencies and corporate media. This call to action resonated with me, especially when considering Snowden's situation. If there is indeed a movement towards restoring free speech and challenging the status quo, could this pave the way for Snowden's return to the United States? Since 2013, Snowden has made a life for himself in Russia. His story has become emblematic of the complex relationship between government secrecy and the public's right to know. The speech I watched, with its bold promises about preventing federal agencies from "colluding" to control information and its calls for investigations into what it termed "online censorship," made me wonder about the possibility of change. The words echoed through the transcript – departments like Homeland Security, Health and Human Services, the FBI, and Department of Justice all mentioned by name. Each reference carried the weight of potential transformation in how our government handles information and those who choose to reveal it. President Trump's emphasis on "reclaiming" free speech struck a particular chord, though the reality of whistleblower protection remains far more complex than any single speech or policy proposal can address. Just as Julian Assange recently returned to Australia, there may be hope for Snowden as well. The political climate appears to be shifting, with more citizens becoming aware of the importance of transparency and the dangers of unchecked power. If the Trump administration truly prioritizes the restoration of free speech, it might also reconsider the stance towards those who have fought for that very principle. Snowden's return would not only symbolize a victory for whistleblowers everywhere but also serve as a reminder that accountability and justice can prevail. It is essential for our democracy to embrace those who have risked everything to expose the truth. As we look towards the future, the light at the end of the tunnel seems brighter. With ongoing conversations about censorship and the rights of individuals, perhaps we are on the brink of a significant shift. It is time to advocate for a system where whistleblowers are protected and celebrated rather than vilified. In conclusion, my hope for Edward Snowden's return is intertwined with the broader movement towards reclaiming free speech and accountability. As we navigate these complex issues, let us remain vigilant and support those who dare to stand up for the truth, for they are the ones who illuminate the path toward a more just society. This piece represents my own personal observations and thoughts on complex issues. Readers should pursue their own research and consult official sources for ground-truthing. Follow me here on Hackernoon for more insights and updates. This morning, after watching President-Elect Donald Trump's announcement regarding his plan to "shatter" the censorship regime, I experienced a profound epiphany. The discussion surrounding free speech and the manipulation of information sparked thoughts about the potential for change in the current political landscape, particularly concerning figures like Edward Snowden. For years, Snowden has been a polarizing figure—celebrated by many as a whistleblower and condemned by others as a traitor. His revelations about government surveillance practices opened a critical dialogue about privacy, security, and the role of government in the lives of citizens. Yet, despite the importance of his contributions to public discourse, he has remained in exile, away from his homeland. In Trump's speech, he emphasized the need to dismantle what he described as a "censorship cartel" created by various entities, including government agencies and corporate media. This call to action resonated with me, especially when considering Snowden's situation. If there is indeed a movement towards restoring free speech and challenging the status quo, could this pave the way for Snowden's return to the United States? Since 2013, Snowden has made a life for himself in Russia. His story has become emblematic of the complex relationship between government secrecy and the public's right to know. The speech I watched, with its bold promises about preventing federal agencies from "colluding" to control information and its calls for investigations into what it termed "online censorship," made me wonder about the possibility of change. The words echoed through the transcript – departments like Homeland Security, Health and Human Services, the FBI, and Department of Justice all mentioned by name. Each reference carried the weight of potential transformation in how our government handles information and those who choose to reveal it. President Trump's emphasis on "reclaiming" free speech struck a particular chord, though the reality of whistleblower protection remains far more complex than any single speech or policy proposal can address. Just as Julian Assange recently returned to Australia, there may be hope for Snowden as well. The political climate appears to be shifting, with more citizens becoming aware of the importance of transparency and the dangers of unchecked power. If the Trump administration truly prioritizes the restoration of free speech, it might also reconsider the stance towards those who have fought for that very principle. Snowden's return would not only symbolize a victory for whistleblowers everywhere but also serve as a reminder that accountability and justice can prevail. It is essential for our democracy to embrace those who have risked everything to expose the truth. As we look towards the future, the light at the end of the tunnel seems brighter. With ongoing conversations about censorship and the rights of individuals, perhaps we are on the brink of a significant shift. It is time to advocate for a system where whistleblowers are protected and celebrated rather than vilified. In conclusion, my hope for Edward Snowden's return is intertwined with the broader movement towards reclaiming free speech and accountability. As we navigate these complex issues, let us remain vigilant and support those who dare to stand up for the truth, for they are the ones who illuminate the path toward a more just society. This piece represents my own personal observations and thoughts on complex issues. Readers should pursue their own research and consult official sources for ground-truthing. Follow me here on Hackernoon for more insights and updates. This piece represents my own personal observations and thoughts on complex issues. Readers should pursue their own research and consult official sources for ground-truthing. Follow me here on Hackernoon for more insights and updates. Hackernoon

Finding Hope: The Possibility of Edward Snowden's Return

আশার খোঁজ: এডওয়ার্ড স্নোডেনের প্রত্যাবর্তনের সম্ভাবনা

Hoffnung finden: Die Möglichkeit einer Rückkehr von Edward Snowden

Encontrando esperanza: la posibilidad del regreso de Edward Snowden

Trouver l'espoir : la possibilité du retour d'Edward Snowden

आशा की तलाश: एडवर्ड स्नोडेन की वापसी की संभावना

希望を見つける：エドワード・スノーデンの復帰の可能性

희망을 찾다: 에드워드 스노든의 귀환 가능성

Encontrando esperança: a possibilidade do retorno de Edward Snowden

В поисках надежды: возможность возвращения Эдварда Сноудена

Umut Bulmak: Edward Snowden'ın Geri Dönme İhtimali

Tìm kiếm hy vọng: Khả năng Edward Snowden trở về

寻找希望：爱德华·斯诺登回归的可能性

The art of prompt engineering is becoming a critical skill for effectively harnessing AI's potential as artificial intelligence continues to evolve. Yet, for all its importance, prompt engineering is often reduced to simplistic advice—tips like “be specific” or quick hacks such as “let’s think step by step.” While helpful for beginners, these pointers barely scratch the surface. True prompt engineering requires a much deeper understanding of the mechanics behind AI responses, moving far beyond what’s often little more than zero-shot Google searching. Here we go through some foundational principles to explain why good prompting actually works. Embeddings: The Foundation of Prompt Understanding One of the foundational concepts in effective prompting is understanding embeddings. When we input a prompt into a model, it doesn’t process language as we do; instead, each word or phrase is transformed into an embedding, a numerical representation that captures both the meaning and context of the language. These embeddings act as the model’s internal mapping, helping it interpret relationships, nuances, and associations between words. For prompt engineers, understanding embeddings is essential for prompt intuition. The way a word is represented in embedding space determines how the model interprets it, affecting the content and focus of responses. For example, if a prompt requests information on “manufacturing,” the model will pull content based on embeddings linked to those terms, potentially drawing from areas like production processes, supply chain logistics, or technological advancements. A prompt engineer can influence these responses by skillfully crafting the wording, improving the relevance and depth of the output. Understanding embeddings also empowers prompt engineers to chain prompts better, where each response builds on previous interactions. By crafting sequences with embedding relationships in mind, prompt engineers create complex, context-aware conversations that adapt to specific needs, producing detailed, responsive outputs. The Role of Attention: Guiding Model Focus Another core component in prompt engineering is attention. Unlike human attention, which can dynamically shift between different topics, AI attention assigns weight to relationships between words within the prompt and response text. Attention helps the model prioritize certain parts of the input, and a well-crafted prompt uses this to shape the model’s focus. Take, for instance, a prompt asking for a nuanced view of a cat sitting on a mat. Prompt engineers can guide the model to emphasize relevant aspects of the prompt's subject, such as the cat's posture, the texture and color of the mat, or the overall scene’s ambiance, leading to a more detailed and vivid response. This approach to prompting. known as directional stimulus, allows the extraction of richer, more relevant answers from the model. RAG: Retrieval-Augmented Generation for Enhanced Responses Retrieval-Augmented Generation (RAG) combines generative AI with retrieval systems to improve the accuracy and relevance of responses. Traditional prompt engineering relies on a model’s internal knowledge, which can be outdated or incomplete. RAG addresses this by adding a retrieval layer that pulls in current, contextually relevant information from external sources like databases or search engines, enabling the model to provide more accurate, up-to-date answers. RAG is transformative for prompt engineering. It allows engineers to structure prompts that direct both the retrieval and generative components to work together, especially valuable for complex or fast-evolving subjects. For instance, a prompt about “recent changes in data privacy laws” could trigger the model to fetch the latest legal documents or news, which it then combines with internal information via RAG to synthesize into a coherent response. This dual capability reduces hallucinations and keeps the model grounded in real-time, reliable information. Prompt engineers working within RAG frameworks can therefore create interactions that are both innovative and anchored in current knowledge, adding an extra layer of sophistication to the AI’s output. Image Prompting and the Diffusion Process In image generation, prompt engineering takes on another layer of complexity. Here, prompts are not just about words but about guiding a model through the diffusion process—a technique where random noise is iteratively transformed into a coherent image. The diffusion process relies on iterative adjustments, and each cycle builds upon the previous one to refine an image based on the prompt's guidance. When crafting prompts for image generation, understanding diffusion is essential. Descriptions like “award winning photography” or “fashion photography” can yield varying results depending on how the model interprets "award winning" or “fashion.” Prompt engineers must understand the model’s interpretation of spatial relationships, color, and composition to produce a desired outcome. It’s about more than aesthetics; it’s about leading the model through each step to align with the intended visual concept. Hallucinations and the Pitfalls of Poor Prompting One of the most challenging aspects of AI prompting is managing hallucinations—when a model produces content that sounds plausible but is entirely fabricated. Hallucinations often result from vague or poorly structured prompts, which leave too much to the model’s interpretation. In fields like healthcare or finance, where accuracy is critical, even a small hallucination can have serious consequences. Skilled prompt engineers prevent this by constructing prompts that minimize ambiguity, verifying outputs with fact-checking prompts, and maintaining strict clarity and focus. Why Prompt Engineering is a Real Skill Ultimately, prompt engineering is a genuine skill that combines language intuition with a deep technical understanding of AI mechanics. Unlike simple search queries, prompt engineers with knowledge of embeddings, attention mechanisms, and model limitations are able to craft better prompts. Mastering these elements enables prompt engineers to shape responses that are precise, reliable, and often surprisingly creative. The need for skilled prompt engineers will only grow as AI technology continues to integrate into industries from education to entertainment. They’ll be essential not just for generating content, but for refining AI applications, troubleshooting outputs, and optimizing model performance across fields. Far from a passing trend, prompt engineering is shaping up to be one of the defining skills of the AI era—an essential tool for building communication between humans and machines. About Me: 25+ year IT veteran combining data, AI, risk management, strategy, and education. 4x global hackathon winner and social impact from data advocate. Currently working to jumpstart the AI workforce in the Philippines. Learn more about me here. The art of prompt engineering is becoming a critical skill for effectively harnessing AI's potential as artificial intelligence continues to evolve. Yet, for all its importance, prompt engineering is often reduced to simplistic advice—tips like “be specific” or quick hacks such as “let’s think step by step.” While helpful for beginners, these pointers barely scratch the surface. True prompt engineering requires a much deeper understanding of the mechanics behind AI responses, moving far beyond what’s often little more than zero-shot Google searching. Here we go through some foundational principles to explain why good prompting actually works. Embeddings: The Foundation of Prompt Understanding One of the foundational concepts in effective prompting is understanding embeddings. When we input a prompt into a model, it doesn’t process language as we do; instead, each word or phrase is transformed into an embedding, a numerical representation that captures both the meaning and context of the language. These embeddings act as the model’s internal mapping, helping it interpret relationships, nuances, and associations between words. a numerical representation For prompt engineers, understanding embeddings is essential for prompt intuition. The way a word is represented in embedding space determines how the model interprets it, affecting the content and focus of responses. For example, if a prompt requests information on “manufacturing,” the model will pull content based on embeddings linked to those terms, potentially drawing from areas like production processes, supply chain logistics, or technological advancements. A prompt engineer can influence these responses by skillfully crafting the wording, improving the relevance and depth of the output. Understanding embeddings also empowers prompt engineers to chain prompts better , where each response builds on previous interactions. By crafting sequences with embedding relationships in mind, prompt engineers create complex, context-aware conversations that adapt to specific needs, producing detailed, responsive outputs. chain prompts better The Role of Attention: Guiding Model Focus Another core component in prompt engineering is attention. Unlike human attention, which can dynamically shift between different topics, AI attention assigns weight to relationships between words within the prompt and response text. Attention helps the model prioritize certain parts of the input, and a well-crafted prompt uses this to shape the model’s focus. Take, for instance, a prompt asking for a nuanced view of a cat sitting on a mat. relationships between words Prompt engineers can guide the model to emphasize relevant aspects of the prompt's subject, such as the cat's posture, the texture and color of the mat, or the overall scene’s ambiance, leading to a more detailed and vivid response. This approach to prompting. known as directional stimulus , allows the extraction of richer, more relevant answers from the model. directional stimulus RAG: Retrieval-Augmented Generation for Enhanced Responses Retrieval-Augmented Generation (RAG) combines generative AI with retrieval systems to improve the accuracy and relevance of responses. Traditional prompt engineering relies on a model’s internal knowledge, which can be outdated or incomplete. RAG addresses this by adding a retrieval layer that pulls in current, contextually relevant information from external sources like databases or search engines, enabling the model to provide more accurate, up-to-date answers. adding a retrieval layer RAG is transformative for prompt engineering. It allows engineers to structure prompts that direct both the retrieval and generative components to work together, especially valuable for complex or fast-evolving subjects. For instance, a prompt about “recent changes in data privacy laws” could trigger the model to fetch the latest legal documents or news, which it then combines with internal information via RAG to synthesize into a coherent response. This dual capability reduces hallucinations and keeps the model grounded in real-time, reliable information. Prompt engineers working within RAG frameworks can therefore create interactions that are both innovative and anchored in current knowledge, adding an extra layer of sophistication to the AI’s output. Image Prompting and the Diffusion Process In image generation, prompt engineering takes on another layer of complexity. Here, prompts are not just about words but about guiding a model through the diffusion process—a technique where random noise is iteratively transformed into a coherent image. The diffusion process relies on iterative adjustments, and each cycle builds upon the previous one to refine an image based on the prompt's guidance. random noise is iteratively transformed When crafting prompts for image generation, understanding diffusion is essential. Descriptions like “award winning photography” or “fashion photography” can yield varying results depending on how the model interprets "award winning" or “fashion.” Prompt engineers must understand the model’s interpretation of spatial relationships, color, and composition to produce a desired outcome. It’s about more than aesthetics; it’s about leading the model through each step to align with the intended visual concept. Hallucinations and the Pitfalls of Poor Prompting One of the most challenging aspects of AI prompting is managing hallucinations—when a model produces content that sounds plausible but is entirely fabricated. Hallucinations often result from vague or poorly structured prompts , which leave too much to the model’s interpretation. In fields like healthcare or finance, where accuracy is critical, even a small hallucination can have serious consequences. vague or poorly structured prompts Skilled prompt engineers prevent this by constructing prompts that minimize ambiguity , verifying outputs with fact-checking prompts, and maintaining strict clarity and focus. prompts that minimize ambiguity Why Prompt Engineering is a Real Skill Ultimately, prompt engineering is a genuine skill that combines language intuition with a deep technical understanding of AI mechanics. Unlike simple search queries, prompt engineers with knowledge of embeddings, attention mechanisms, and model limitations are able to craft better prompts. Mastering these elements enables prompt engineers to shape responses that are precise, reliable, and often surprisingly creative. The need for skilled prompt engineers will only grow as AI technology continues to integrate into industries from education to entertainment. They’ll be essential not just for generating content, but for refining AI applications, troubleshooting outputs, and optimizing model performance across fields. Far from a passing trend, prompt engineering is shaping up to be one of the defining skills of the AI era—an essential tool for building communication between humans and machines. About Me: 25+ year IT veteran combining data, AI, risk management, strategy, and education. 4x global hackathon winner and social impact from data advocate. Currently working to jumpstart the AI workforce in the Philippines. Learn more about me here . here

Explaining Prompt Engineering

প্রম্পট ইঞ্জিনিয়ারিং ব্যাখ্যা করা

Erklärung zu Prompt Engineering

Explicando la ingeniería rápida

Explication de l'ingénierie rapide

प्रॉम्प्ट इंजीनियरिंग की व्याख्या

プロンプトエンジニアリングの説明

프롬프트 엔지니어링 설명

Explicando a Engenharia Prompt

Объяснение оперативной инженерии

İstemli Mühendisliğin Açıklanması

Giải thích về Kỹ thuật Prompt

解释快速工程

How Society is Failing Young Professionals: Workplace Stress and the Mental Health Crisis

Why AI-Powered Image Hosting APIs Are Essential for Modern Businesses

Waarom AI-aangedrewe beeldgasheer-API's noodsaaklik is vir moderne besighede

Proč jsou rozhraní API pro hostování obrázků s umělou inteligencí nezbytná pro moderní podniky

Warum KI-gestützte Bildhosting-APIs für moderne Unternehmen unverzichtbar sind

Por qué las API de alojamiento de imágenes basadas en IA son esenciales para las empresas modernas

Miért nélkülözhetetlenek az AI-alapú képtárhely API-k a modern vállalkozások számára?

AI を活用した画像ホスティング API が現代のビジネスに不可欠な理由

Неліктен AI-мен жұмыс істейтін кескін хостинг API интерфейстері қазіргі заманғы бизнес үшін өте маңызды?

Kāpēc ar AI darbināmas attēlu mitināšanas API ir būtiskas mūsdienu uzņēmumiem

Ke ka baka la’ng di-API tša go Amogela Diswantšho tšeo di Matlafatšwago ke AI di le Bohlokwa bakeng sa Dikgwebo tša Sebjalebjale

Почему API-интерфейсы хостинга изображений на базе искусственного интеллекта необходимы современному бизнесу

නවීන ව්‍යාපාර සඳහා AI-බලවත් රූප සත්කාරක API අත්‍යවශ්‍ය වන්නේ ඇයි?

Tại sao API lưu trữ hình ảnh hỗ trợ AI lại cần thiết cho các doanh nghiệp hiện đại

In the bull market, when BTC goes up and breaks its all-time high (ATH), the accumulation process for a new ATH is underway. This is good for altcoins, which go up next. SOL and ETH are the first altcoins to start moving, just like many other Layer 1’s. Investors are looking for ways to maximize their returns on ETH because of gas fees and exploring alternative options, which brings us to Layer 2. Projects like BASE & ARBITRIUM, DOGECHAIN, OPTIMISM, POLYGON, and others that are built on ETH are starting to grow too. Once these altcoins start to gain traction, investors start looking for ways to diversify their profits. Things started to change this year. Once ETFs were approved for BTC, altcoins didn't see much growth. They rarely took off. While some meme tokens saw huge gains (up to 10x), altcoins tried to triple their value and then fell back to their original levels. The same thing happened again after the second big spike, this time with an ETF for ETH. Once again, while ETH peaked, altcoins stayed at the bottom. And yes, meme tokens were making waves again. Crypto wallet activity is at an all-time high, and it is still on the rise. Over the course of the year, it has tripled, compared to previous cycles that saw a tenfold increase. I'm seeing the greatest increase in wallet activity on Solana and Base. This suggests that the alt season hasn't started yet, and could indirectly confirm the BTC supercycle theory. The rise of decentralized launchpads has been a major driver of the growth of these networks. You don't need to know any coding to create a token. But launchpads have some technical limitations, they're great for creating template tokens. It's worth noting that the only category of template tokens that can attract a large community are meme tokens! That's why the most successful and first Launchpads is based on Solana and was designed to run meme tokens. Solana is where a lot of memes started. Is that just a coincidence? It's not a coincidence. It's a trend. Many memes launched on pump.fun are later launched as analogs on other networks, where some of them become more successful than the original projects. The second most popular moonshot launchpad from dexscreener also started on Solana and now supports Base. ETH saw little change in developers’ activity from year to year (though it's still the leader in this area). BTC and its L2s, on the other hand, have seen a significant increase in development activity, doubling in the past year. This growth is likely to continue and set new trends in the coming year. To avoid confusion, I will refer to the new Meme tokens currently trending in the short-term market as "Memes 2.0". There are several differences between Memes 2.0 and Meme tokens from previous cycles. The main difference is how the tokens were launched. Memes 2.0 launch on launchpads using the Bonding Curve, which we call a fair launch. What makes this launch different is how the tokens are distributed after they're listed on the dex. Specifically, only a small percentage of the supply is available for purchase in the liquidity pool. This means that if you try to sell a large number of tokens, the price will drop dramatically. In fact, 99.999% of these tokens going to 0 after being listed on dex. This is an important point. I like to think of this as natural selection. If meme 2.0 has a strong narrative and a healthy community by the time when whales and/or developers are out, a CTO (community token owned) is created. The community takes ownership of the project's development. With a strong community, the project is already more likely to succeed than at the Bonding Curve level. A short-term speculative meme 2.0 can become a long-term meme. So the main things that differentiate memes 2.0 from those of previous cycles are how popular the CTOs are, how the tokens are distributed, and how they're launched. Some good examples are GOAT, Michi, FWOG, and MOODENG. On top of that, memes 2.0 have become a kind of tokenized representation of everything that happens in the world. It's like a news channel. If you want to see what's trending, just go to pump.fun and look at the top active tokens :) Murad's theory of the memcoin supercycle suggests that we may soon see a cycle in which memcoins surpass traditional coins. This will be based on community sentiment, emotion, and the speculative lure of getting rich quickly. Memcoins are different from traditional projects because they grow through hype and the ability to rally large communities. The main things to watch out for are: The popularity of meme launchpads (pump.fun, moonshot) and the fact that you can deploy memtokens almost for free without any programming skills or blockchain knowledge.
Many moonboys on X and other socials
More and more people are looking to invest in high risk assets like memcoins due to the instability in the world and high inflation across the board
We're seeing a shift where communities driven by profit are replacing those driven by blockchain and technology.
Major exchanges like Binance and Coinbase are listing more and more meme tokens, fueling a speculative frenzy. Meme token supercycle theory has no real fundamentals like memcoins itself. They attract traders with the promise of quick profits. There's no real utility or technology behind them. It's just people chasing the dream of becoming a millionaire with just $10. It reminds me of the ICO boom I was a part of. Even the most obvious scams raised money. They had fake profiles of team members, renamed social networks, made beautiful websites, and that was it. They just sold the standard tokens and out. Yes, there were some real projects, but only a few. Investors were just as fixated on quick profits as they are today. After the whole system crashed, most of them lost their money. Then they started thinking about technology and utility. It seems that the new generation of investors hasn’t learned from our mistakes and is falling into the same traps we did. Memcoins like Dogecoin, Pepe, and Shiba Inu will probably benefit from this hype around memes either way. On the other hand, most Meme 2.0 will experience a rug pool, not a supercycle:) Given the current market conditions, it's likely that the surviving 2.0 memtokens will see zero value due to their low liquidity and high FDV. References State of Crypto Report 2024
The case for a 2024 bull run In the bull market, when BTC goes up and breaks its all-time high (ATH), the accumulation process for a new ATH is underway. This is good for altcoins, which go up next. SOL and ETH are the first altcoins to start moving, just like many other Layer 1’s. Investors are looking for ways to maximize their returns on ETH because of gas fees and exploring alternative options, which brings us to Layer 2. Projects like BASE & ARBITRIUM, DOGECHAIN, OPTIMISM, POLYGON, and others that are built on ETH are starting to grow too. Once these altcoins start to gain traction, investors start looking for ways to diversify their profits. Things started to change this year. Once ETFs were approved for BTC , altcoins didn't see much growth. They rarely took off. While some meme tokens saw huge gains (up to 10x), altcoins tried to triple their value and then fell back to their original levels. The same thing happened again after the second big spike, this time with an ETF for ETH. Once again, while ETH peaked, altcoins stayed at the bottom. And yes, meme tokens were making waves again. BTC Crypto wallet activity is at an all-time high, and it is still on the rise. Over the course of the year, it has tripled, compared to previous cycles that saw a tenfold increase. I'm seeing the greatest increase in wallet activity on Solana and Base. This suggests that the alt season hasn't started yet, and could indirectly confirm the BTC supercycle theory. The rise of decentralized launchpads has been a major driver of the growth of these networks. You don't need to know any coding to create a token. But launchpads have some technical limitations, they're great for creating template tokens. It's worth noting that the only category of template tokens that can attract a large community are meme tokens! That's why the most successful and first Launchpads is based on Solana and was designed to run meme tokens. Solana is where a lot of memes started. Is that just a coincidence? It's not a coincidence. It's a trend. Many memes launched on pump.fun are later launched as analogs on other networks, where some of them become more successful than the original projects. The second most popular moonshot launchpad from dexscreener also started on Solana and now supports Base. ETH saw little change in developers’ activity from year to year (though it's still the leader in this area). BTC and its L2s, on the other hand, have seen a significant increase in development activity, doubling in the past year. This growth is likely to continue and set new trends in the coming year. To avoid confusion, I will refer to the new Meme tokens currently trending in the short-term market as "Memes 2.0". There are several differences between Memes 2.0 and Meme tokens from previous cycles. The main difference is how the tokens were launched. Memes 2.0 launch on launchpads using the Bonding Curve, which we call a fair launch. What makes this launch different is how the tokens are distributed after they're listed on the dex. Specifically, only a small percentage of the supply is available for purchase in the liquidity pool. This means that if you try to sell a large number of tokens, the price will drop dramatically. In fact, 99.999% of these tokens going to 0 after being listed on dex. This is an important point. I like to think of this as natural selection. If meme 2.0 has a strong narrative and a healthy community by the time when whales and/or developers are out, a CTO (community token owned) is created. The community takes ownership of the project's development. With a strong community, the project is already more likely to succeed than at the Bonding Curve level. A short-term speculative meme 2.0 can become a long-term meme. So the main things that differentiate memes 2.0 from those of previous cycles are how popular the CTOs are, how the tokens are distributed, and how they're launched. Some good examples are GOAT, Michi, FWOG, and MOODENG. On top of that, memes 2.0 have become a kind of tokenized representation of everything that happens in the world. It's like a news channel. If you want to see what's trending, just go to pump.fun and look at the top active tokens :) Murad's theory of the memcoin supercycle suggests that we may soon see a cycle in which memcoins surpass traditional coins. This will be based on community sentiment, emotion, and the speculative lure of getting rich quickly. Memcoins are different from traditional projects because they grow through hype and the ability to rally large communities. The main things to watch out for are: The popularity of meme launchpads (pump.fun, moonshot) and the fact that you can deploy memtokens almost for free without any programming skills or blockchain knowledge. Many moonboys on X and other socials More and more people are looking to invest in high risk assets like memcoins due to the instability in the world and high inflation across the board We're seeing a shift where communities driven by profit are replacing those driven by blockchain and technology. Major exchanges like Binance and Coinbase are listing more and more meme tokens, fueling a speculative frenzy. The popularity of meme launchpads (pump.fun, moonshot) and the fact that you can deploy memtokens almost for free without any programming skills or blockchain knowledge. Many moonboys on X and other socials More and more people are looking to invest in high risk assets like memcoins due to the instability in the world and high inflation across the board We're seeing a shift where communities driven by profit are replacing those driven by blockchain and technology. Major exchanges like Binance and Coinbase are listing more and more meme tokens, fueling a speculative frenzy. Meme token supercycle theory has no real fundamentals like memcoins itself. They attract traders with the promise of quick profits. There's no real utility or technology behind them. It's just people chasing the dream of becoming a millionaire with just $10. It reminds me of the ICO boom I was a part of. Even the most obvious scams raised money. They had fake profiles of team members, renamed social networks, made beautiful websites, and that was it. They just sold the standard tokens and out. Yes, there were some real projects, but only a few. Investors were just as fixated on quick profits as they are today. After the whole system crashed, most of them lost their money. Then they started thinking about technology and utility. It seems that the new generation of investors hasn’t learned from our mistakes and is falling into the same traps we did. Memcoins like Dogecoin, Pepe, and Shiba Inu will probably benefit from this hype around memes either way. On the other hand, most Meme 2.0 will experience a rug pool, not a supercycle:) Given the current market conditions, it's likely that the surviving 2.0 memtokens will see zero value due to their low liquidity and high FDV. References State of Crypto Report 2024 The case for a 2024 bull run State of Crypto Report 2024 State of Crypto Report 2024 The case for a 2024 bull run The case for a 2024 bull run

Meme Tokens 2.0 or ICO 2.0?

Managing our work lives with digital tools? Easy. Productivity apps? Everywhere. But managing our personal lives? That was a whole different story. Between grocery lists jotted on fridge magnets, sticky notes left on tables, and plans discussed but quickly forgotten, our personal life often felt like a chaotic whirlwind. Despite both of us being heavy users of Notion for work (me as a product manager and her as a PhD student), and paying for Notion AI as part of my job, our personal organization was still…well, messy. We’ve long been fans of Notion for managing our relationship and even went viral for it once, with our blog being featured in places like Morning Brew. But we realized that Notion AI might be the missing piece to elevate what we’d already built in our relationship journal. So, we decided to go into an exploratory phase, wondering: What problems in our daily lives could Notion AI solve for us as a couple? The result? Let’s just say it’s drastically changed our lives. Here are the top three ways Notion AI has made our relationship journal more powerful, efficient, and (dare I say) fun. 1. Date Nights: No More “What Do You Want to Do?” Loops If you’ve ever been stuck in the infinite question loop of “What do you want to do tonight?”, you’ll understand the frustration of indecisive date nights. This was a regular occurrence for us, especially on Friday evenings. We decided to bring Notion AI into the mix. Using our journal, where we already tracked favorite date nights and NYC activities, we built a Date Night Brainstorm page. The page uses a prompt to consider our preferences, past activities, and even restaurant recommendations nearby to suggest an itinerary. Does it always nail it? Not quite. About 70-80% of the time, it generates great suggestions—sometimes even creative ideas we wouldn’t have thought of ourselves. And while it’s not perfect (no AI is), it takes the burden of planning away. Now, instead of endless discussions, we have a ready-made itinerary waiting for us, making date nights smooth, exciting, and much less stressful. 2. Trip Planning: Our New Travel Assistant Trip planning used to be a big deal in our relationship. My partner is known for meticulously detailed itineraries, but now we joke that Notion AI has taken her job. We built two key tools in our template: a Trip Brainstorming page and a Trip Planning page. The Brainstorming page asks us questions like: How many days are we traveling?
What kind of trip are we feeling—outdoorsy, cultural, or relaxed?
Do we have specific destinations or regions in mind? Notion AI then processes our inputs and suggests five or six destinations tailored to our preferences. It’s incredibly helpful for narrowing down options. The Planning page is where the magic happens. Based on our chosen destination and travel dates, Notion AI: Generates a detailed daily itinerary, down to the hour.
Researches the top destinations and activities in the area.
Creates a customized packing list, categorized by type and quantity.
Outputs a checklist of next steps, such as bookings, reservations, and tickets. It’s taken the mental load of planning off our shoulders and made trips smoother than ever. While we joke about how it’s replaced my partner’s role as the “itinerary queen,” it’s also made travel a lot more fun and collaborative for us. 3. Household Management: Meal Planning Made Easy When we first created our relationship journal in 2020, we were newly dating, focused on things like defining our values and planning fun date nights. Fast-forward to today, we’re married, and our lives now involve managing a home together. Enter the Household Page, the newest addition to our journal, and the Notion AI feature that’s been a game-changer: the Recipe Ideation page. Here’s how it works: We input things like: Ingredients we have at home
Calorie and macro targets
Number of servings
Dietary restrictions or preferences Notion AI generates a list of recipes complete with ingredients and instructions. These recipes feed into a Recipe Book that we maintain, making Sunday meal planning a breeze. We tag recipes to specific days of the week, eliminating the stress of deciding what to cook during our busy workweeks. This small change has made our lives significantly more productive and keeps us on track with our health goals. Other Fun Ways We Use Notion AI While these three pages have had the biggest impact, we’ve also added some smaller, fun features that we enjoy daily: Netflix Recommendations based on what we’ve previously watched.
Book and Podcast Suggestions tailored to our shared interests.
A daily Question of the Day to spark meaningful or humorous conversations. These little touches have added even more joy to our routine. Wrapping It Up Notion AI has fundamentally changed how we manage our life together. By taking the mental load off tasks like trip planning, meal prep, and even date night brainstorming, it’s given us more time and space to focus on what matters most: each other. That said, it’s worth mentioning that Notion AI costs an additional $10 per month. If you already use Notion heavily in your workflow, it’s an easy recommendation—it seamlessly enhances what you’re already doing. But it’s definitely a consideration if you’re thinking of adding yet another subscription to your life. If you’re already using Notion AI or curious about its potential, I’d encourage you to try maintaining a relationship journal like ours. It’s a small but transformative way to bring clarity, organization, and intentionality to your personal life. And if you’re interested in trying out the template we’ve built, feel free to check it out. If it doesn’t work for you, we’ll be happy to issue a refund—no hard feelings. Thank you for reading and for supporting us through these years. Here’s to growing and thriving together. Cheers! Managing our work lives with digital tools? Easy. Productivity apps? Everywhere. But managing our personal lives? That was a whole different story. Between grocery lists jotted on fridge magnets, sticky notes left on tables, and plans discussed but quickly forgotten, our personal life often felt like a chaotic whirlwind. Despite both of us being heavy users of Notion for work (me as a product manager and her as a PhD student), and paying for Notion AI as part of my job, our personal organization was still…well, messy. We’ve long been fans of Notion for managing our relationship and even went viral for it once, with our blog being featured in places like Morning Brew . But we realized that Notion AI might be the missing piece to elevate what we’d already built in our relationship journal. So, we decided to go into an exploratory phase, wondering: Morning Brew What problems in our daily lives could Notion AI solve for us as a couple? What problems in our daily lives could Notion AI solve for us as a couple? The result? Let’s just say it’s drastically changed our lives. Here are the top three ways Notion AI has made our relationship journal more powerful, efficient, and (dare I say) fun. 1. Date Nights: No More “What Do You Want to Do?” Loops If you’ve ever been stuck in the infinite question loop of “What do you want to do tonight?”, you’ll understand the frustration of indecisive date nights. This was a regular occurrence for us, especially on Friday evenings. We decided to bring Notion AI into the mix. Using our journal, where we already tracked favorite date nights and NYC activities, we built a Date Night Brainstorm page. The page uses a prompt to consider our preferences, past activities, and even restaurant recommendations nearby to suggest an itinerary. Does it always nail it? Not quite. About 70-80% of the time, it generates great suggestions—sometimes even creative ideas we wouldn’t have thought of ourselves. And while it’s not perfect (no AI is), it takes the burden of planning away. Now, instead of endless discussions, we have a ready-made itinerary waiting for us, making date nights smooth, exciting, and much less stressful. 2. Trip Planning: Our New Travel Assistant Trip planning used to be a big deal in our relationship. My partner is known for meticulously detailed itineraries, but now we joke that Notion AI has taken her job. We built two key tools in our template: a Trip Brainstorming page and a Trip Planning page. The Brainstorming page asks us questions like: Trip Brainstorming Trip Planning Brainstorming How many days are we traveling? What kind of trip are we feeling—outdoorsy, cultural, or relaxed? Do we have specific destinations or regions in mind? How many days are we traveling? What kind of trip are we feeling—outdoorsy, cultural, or relaxed? Do we have specific destinations or regions in mind? Notion AI then processes our inputs and suggests five or six destinations tailored to our preferences. It’s incredibly helpful for narrowing down options. The Planning page is where the magic happens. Based on our chosen destination and travel dates, Notion AI: Planning Generates a detailed daily itinerary, down to the hour. Researches the top destinations and activities in the area. Creates a customized packing list, categorized by type and quantity. Outputs a checklist of next steps, such as bookings, reservations, and tickets. Generates a detailed daily itinerary, down to the hour. Researches the top destinations and activities in the area. Creates a customized packing list, categorized by type and quantity. Outputs a checklist of next steps, such as bookings, reservations, and tickets. It’s taken the mental load of planning off our shoulders and made trips smoother than ever. While we joke about how it’s replaced my partner’s role as the “itinerary queen,” it’s also made travel a lot more fun and collaborative for us. 3. Household Management: Meal Planning Made Easy When we first created our relationship journal in 2020, we were newly dating, focused on things like defining our values and planning fun date nights. Fast-forward to today, we’re married, and our lives now involve managing a home together. Enter the Household Page, the newest addition to our journal, and the Notion AI feature that’s been a game-changer: the Recipe Ideation page. Here’s how it works: We input things like: Ingredients we have at home Calorie and macro targets Number of servings Dietary restrictions or preferences Ingredients we have at home Calorie and macro targets Number of servings Dietary restrictions or preferences Notion AI generates a list of recipes complete with ingredients and instructions. These recipes feed into a Recipe Book that we maintain, making Sunday meal planning a breeze. We tag recipes to specific days of the week, eliminating the stress of deciding what to cook during our busy workweeks. This small change has made our lives significantly more productive and keeps us on track with our health goals. Other Fun Ways We Use Notion AI While these three pages have had the biggest impact, we’ve also added some smaller, fun features that we enjoy daily: Netflix Recommendations based on what we’ve previously watched. Book and Podcast Suggestions tailored to our shared interests. A daily Question of the Day to spark meaningful or humorous conversations. Netflix Recommendations based on what we’ve previously watched. Book and Podcast Suggestions tailored to our shared interests. A daily Question of the Day to spark meaningful or humorous conversations. These little touches have added even more joy to our routine. Wrapping It Up Notion AI has fundamentally changed how we manage our life together. By taking the mental load off tasks like trip planning, meal prep, and even date night brainstorming, it’s given us more time and space to focus on what matters most: each other. That said, it’s worth mentioning that Notion AI costs an additional $10 per month. If you already use Notion heavily in your workflow, it’s an easy recommendation—it seamlessly enhances what you’re already doing. But it’s definitely a consideration if you’re thinking of adding yet another subscription to your life. If you’re already using Notion AI or curious about its potential, I’d encourage you to try maintaining a relationship journal like ours. It’s a small but transformative way to bring clarity, organization, and intentionality to your personal life. And if you’re interested in trying out the template we’ve built, feel free to check it out . If it doesn’t work for you, we’ll be happy to issue a refund—no hard feelings. check it out Thank you for reading and for supporting us through these years. Here’s to growing and thriving together. Cheers!

We Revolutionized Managing Our Life as a Couple with AI and Notion. Here’s What We Did.

Stories, News & Mentions about J2HEALTH

J2health’s Stories on HackerNoon