Exploding moons, nature turning against humanity, and superhuman secret agents are all possible in the realm of this exciting genre. Our list of the top seven best action books is chock-full of titles brimming with thrilling adventure. “Journey to the Ce ...

Best Action Books: Top 7 Adventure Novels Most Recommended By Experts

For instance, Pitt possesses telepathic abilities in addition to his physical prowess, toughness, berserker rage, claws, superhuman speed, and accelerated healing, to name a few. He also has a power called "Emotional Control," using which Pitt can disturb ...

10 most powerful superheroes in the Image Comics universe

Since the introduction of films, horror has always been something viewers have resonated with. With the first horror film ever, “ Le Manoir du Diable,” being released in 1896,

Who is the most iconic horror antagonist?

Wolverine is a two-cost, two-Power card with the ability that reads “When this is discarded or destroyed, regenerate it at a random location with +2 Power.” In terms of stats, Wolverine may not be your best choice, but when his ability is fully maximized by creating decks that aim to destroy or discard him, he can be a potential game-changer.

Best Wolverine decks in Marvel Snap (November 2023)

In the wake of OpenAI's latest announcement, AI regulation is finally starting to shape up.

Biden signs sweeping new executive order on the heels of OpenAI's latest big announcement

Luke Cage, with his superhuman strength and bulletproof skin, is the perfect hero to take down the vampires plaguing New York City, and his resourcefulness could potentially make him an even better vampire hunter than Blade. Vampires have become a common ...

Marvel's Luke Cage is Perfect for Fighting Vampires

Luke Cage, with his superhuman strength and bulletproof skin, is the perfect hero to take down the vampires plaguing New York City, and his resourcefulness could potentially make him an even better vampire hunter than Blade.

While Marvel Comics are jam-packed with powerful, superhuman, and even cosmic heroes and villains, they also feature a great many characters who are just as notable for their smarts. Many of these characters,

16 Smartest Marvel Comics Characters, Ranked

Aaron Rodgers was dropping back on the field at MetLife Stadium on Sunday. He is VERY far ahead of the standard Achilles recovery timeline!

Let’s Break Down Aaron Rodgers’ Superhuman Recovery After New Video Of Drop-Back Throws

Having come a long way since her initial brash attitude as a soldier, does Mikasa's maturity and strength outmatch Levi in his injured state?

Attack On Titan: Is Adult Mikasa Stronger Than Levi?

Superhuman

superhuman

Productivity software is indispensable for all of us. Dedicated tools for creating spreadsheets, writing down notes, managing our to-dos etc. ensure that we are productive in our professions. Incumbent vendors (e.g. Microsoft, Google) have dominated this space for many years — yet, there is . New challengers such as the E-Mail client or the note-taking tool can compete by developing highly opinionated products which deliberately omit flexibility. Instead, the software is stringently crafted around a very particular way of approaching a process, which is inherently seen as superior by the firm. a chance for new players to enter the stage Superhuman Roam Research 1st and 2nd wave of differentiation Since the emergence of productivity software in the 1980s, it was sufficient for a long time to entirely focus on features ( ). This is not surprising as a large value could already be created by adding desired features to very rudimentary software products. Microsoft, with its office suite, has emerged clearly as the winner of this era. 1st wave of differentiation From the 2000s onwards, cloud computing, especially SaaS as the new delivery model had a significant impact on the productivity space (enabled by the explosion of internet speed). Advantages such as low setup & infrastructure costs, high scalability and automatic updates have been so revolutionary ( ) that on-premise software has become (almost) obsolete. The success of Salesforce or Google’s G Suite is attributed to leading the way in offering productivity software as a service. 2nd wave of differentiation It is very challenging for a young company to compete on features with the large incumbents of the productivity space. Furthermore, SaaS has established itself as the default. So how can startups successfully differentiate? 3rd wave of differentiation = opinionated software Your average e-mail client, browser software, spreadsheet software does not have an regarding how it should be used. Instead, it provides a maximum of flexibility which is appreciated by the average user, yet, not optimal for particular target groups. This enables an untapped potential for challenger firms which are capable of two things ( ): opinion 3rd wave of differentiation Break problems for a very specific target group of users down into core pieces (see ) first principles thinking Elaborate a firm opinion which UX represents the solution (optimize function rather than form) and rigorously execute on it. They say software should always be as flexible as possible. We think that’s bullshit. The best software has a vision. The best software takes sides. When someone uses software, they’re not just looking for features, they’re looking for an approach. Make Opinionated Software (Basecamp) Let’s focus on Superhuman and Roam Research. Two firms which have demonstrated what an outstanding opinionated productivity software product can look like. Superhuman is an e-mail client (mobile/desktop) for a target group of users who regularly receive a high number of e-mails. The company has received $33M in funding from prestigious investors such as a16z and First Round Capital. The firm has been able to create a willingness to pay of $30 per user per month for its premium e-mail client. Superhuman They need 300k users for accessing “unicorn area” (assuming a target of $100M ARR). Currently, there are almost . This trajectory (founded in 2014) becomes even more impressive after realizing that there are only 50 people working for the firm (according to ). 300k on the waitlist Linkedin Rahul Vohra, the CEO & Co-Founder, has a clear stance on the principles behind their software product. Superhuman’s value proposition of “creating the fastest e-mail experience in the world” can only be achieved by building the product . It is not about adding elements of gamification to a well-thought-out e-mail software. from the ground up on game design principles Instead, it is about crafting an entirely new e-mail experience based on game design principles (game design ≠ gamification). Based on these principles, Superhuman creates a user experience which stringently guides you through the path to inbox zero. For instance, after having replied to an e-mail, you are (without any distraction) directed to the next e-mail. There is no required “halt in-between” at your inbox (distraction) from where you manually select the next e-mail — highly opinionated on how you should go through your e-mails. My favorite example of the underlying game design principles is related to the . This is the mental state in which a person performing an activity is fully immersed in a feeling of energized focus, full involvement, and enjoyment in the process of the activity. flow state The Experience Fluctuation Model (see below) reasons that flow is achieved when we Therefore, Superhuman deliberately increases the challenge level by nudging you to hit inbox zero without touching the mouse — keyboard shortcuts play a crucial role in the Superhuman-UX. balance high perceived skill with high perceived challenge. Interestingly, this has inspired already many other software startups leveraging “shortcuts to superpowers” such as , or . Linear Height Quill Roam Research Note-taking software is highly associated with Evernote. A company which became in 2012/2013 so hyped that and it started selling physical products (e.g. ) before it became known for going through . (funded by Global Village and Chapter One Ventures) puts note-taking back into the spotlight — this time from a different angle. applicants crafted songs over-priced business socks several crises Roam Research Conor White-Sullivan, the CEO & founder of Roam Research has a firm opinion. He is convinced that the common hierarchical approach (main element→ sub element→ note), an approach every popular note-taking solutions (incl. Evernote & Notion) is based on, is dysfunctional for many use cases because it does not support the way our brain works. We do not think in a structured and hierarchical manner. Instead, we think multi-dimensional and associate various pieces of information with each other. In Conor’s view, by storing knowledge in an interlinked manner. This is precisely how Roam Research’s powerful bi-directional linking feature works. the software should mimic how our brain works By merely adding double square brackets around a term “[[xxx]]”, a separate note is automatically created and conveniently linked both ways. The product is highly inspired by the , invented by the renowned German sociologist . His well-thought-out of 90,000 interlinked notes enabled him to build a web of knowledge, retrieve the right memory when needed, find meaningful connections among topics and develop ideas or arguments. Zettelkasten methodology Niklas Luhmann system The system becomes our 2nd brain and enables us to collaborate with our past and future self. Thanks to the Zettelkasten, Niklas Luhmann was able to establish himself as one of the most productive scientists — he published more than . 70(!) books and 500 scholarly articles Roam Research is a highly opinionated product in the sense that it takes a clear side for the Zettelkasten approach. And it works very well. Roam Research has inspired cult-like devotion from early users (#Roamcult). After going into public beta in October 2019, Twitter is full of users sharing their passion and how they are using it. The company is still in its early days, and it is hard to project if the team can expand beyond a particular niche (currently: people who think about how they think) and defend itself against copycats. Executing on their vision of enabling users to share their notes (=knowledge) with others might be the key to success as it creates defensibility through network effects. It seems that 2020 is the year a hype on note-taking is back — this time it might live up to its promises. Opinionated productivity software as a great chance Developing successful productivity software is hard. Incumbents with deep pockets and top-notch product/tech skills can rely on strong network effects as a mean of extensibility. This effect is stronger for some types of productivity software (e.g. collaboration/communication) than for others (e.g. note-taking). Despite these challenges, I’m optimistic that the 3rd wave of differentiation represents an excellent opportunity for businesses. This is mainly due to 4 reasons: 💰 By building opinionated software, you deliberately leave a large share of the market unexploited — it’s not a “lowest common denominator” approach. Fortunately, the overall category of productivity software represents a very large market (mainly because it’s industry-agnostic). Therefore, even a slice of the market is often sufficient. Large market size : 💳: Superhuman costs $30 per month. Roam plans to charge $15 per month — that’s an unprecedented willingness to pay. Will this scare people away? Definitely, but this is part of the plan (see above). A new willingness to pay The first product/solution a company is building shapes its culture as well as its product pipeline for a very long time — certain boundaries are set. Therefore, there’s ample space for startups/challengers to come up with novel opinionated software ideas that an incumbent won’t be able to tackle. Leapfrogging incumbents 🥇: 👨🏻‍🎤: Products such as Superhuman or Roam Research stand for something — they polarize. An E-Mail ending with “sent via Superhuman” signals professionalism. It’s more than the aura of a new product — it can become a sustainable branding advantage. Overall, I’m convinced that the importance of branding will increase in the software space. Opinionated = Strong character Further opportunities — what’s next? I’m excited to see which further companies will emerge in the productivity software space. Some examples coming into my head: The user experience hasn’t changed for many years. Products barely differ. is on it — who else? Web browsing: The Browser Company More distraction by more tools/media/etc. which constantly surround us. WFH blurs the lines between work & free time even further, thereby only aggravating the issue. An anti-procrastination/pro-flow tool: Excel/Google Sheets do not guide us effectively through processes around spreadsheets. Certain use cases represent interesting opportunities (e.g. for finance professionals) Software on top of spreadsheets: Layer There have been many examples in the consumer space demonstrating that highly opinionated products can lead to tremendous success (e.g. Twitter’s 280 character limit, Snapchat’s disappearing messages). It’s time to see more of this in the productivity software space. If you liked this article, I’m happy to connect and exchange thoughts ( / ). Twitter Linkedin Previously published at https://medium.com/@felixschmitt/the-era-of-opinionated-productivity-software-superhuman-roam-whats-next-3d454e28312f

The Era of Opinionated Productivity Software: Superhuman, Roam, What’s Next?

A.I. will eat the world, but probably not entrepreneurship itself That A.I. is able to beat humans in board games is common knowledge by now. How that is relevant to companies is less clear. Normally scientists could be of help here. Clearing up the mess by debunking the hype and showing how trustworthy the really is. But many of the good A.I. researchers are connected to big tech companies and these companies have little incentive to a hype when that negatively impacts their share holder value. technology debunk As I am not yet hired by one of the big tech companies (you do have my email address, right?) I will do some clearing up for you. In this post I will argue that: Executives should embrace the strategic advantages of superhuman A.I. technologies, knowing what can be automated and what not Companies should reposition there human workforce to focus on tasks that require supercomputer qualities Algorithms can already automate creative tasks (yikes!) I’ll jump right in it. Last year , a master student of graduated who was wondering whether you can apply data science to a creative task. He did research into the (semi-)automated creation of advertisement product compositions. Let that sink in for a moment… Kay Hoogland our institute, That’s right. He used an algorithm to , to _d_esign advertisements. create . Advertisement 1 and 3 are computer generated. Kay Hoogland at PyData 2018 True, the advertisements are not that complex. But a creative task, normally done by higly paid professional designers, was automated. In the limited time of his master thesis he made an algorithm that did a better job designing advertisements than designers themselves in over 50% of the cases. Let that sink in again… The resulting application co-creates compositions with graphic designers by making use of an evolutionary algorithm with user input. So the current solution is to still involve designers in the decision making process of which advertisement design works best. This last task could be easily automated as well, if you would let actual click-rate by consumers determine what design works best. Sink about that. Now here is the thing. Even though an algorithm does the job better than any human, it will not the job. get “Even though an algorithm does the job better than any human, it will not get the job” This lack of trust in computers and what they do is a known issue. Already since the 50s we know that . This means that the implicit expertise of a doctor is no match for a model who uses all available data. But doctors, or humans more generally, do not trust most jobs to computers. statistical computer predictions outperform clinical human predictions Humans think their own expertise is somehow more special. While all the data shows it is not. Here is an unfair competitive advantage if I ever seen one: Trust data, trust algorithms and outperform any competitor, that still relies on human decision making, with ease. Whatever you do, please start today with embracing A.I. Go on, give it a hug 🤗. From my workshop Designing Intelligent Disruption The possibilities of A.I. seem endless. They are, of course, not. The not so endless possibilities of A.I. A.I. seems to automate playing board games quite good. Within a context of set rules and a very simple world A.I. is winning hard. But how does that translate to business? Boring tasks seem to be automated rather well. Even simple design and copywriting tasks. But as I have mentioned earlier, we are far away from o . A.I. doesn’t get nuance or multiple intrepretations of a situation. There is still an uncanny valley between humans and A.I. Another fundamental problem is understanding causal relations, although it sees patterns and relations in any data, it can’t yet figure out . vercoming the barrier of meaning cause and effect How speedy the developments will really go is uncertain. And even though the earlier mentioned big tech A.I. researchers are good at creating about fundamental research issues online, these issues will not be tackled any time soon. Until that time it is important to develop a human capital strategy accordingly. a big show “use real humans for new business & A.I. for existing business” I began this year with a plea for using to create long term value for your company. I am extending that plea with using real humans for new business and A.I. for existing and well known business. new data instead of existing data Being an entrepreneur, creating entirely new business; linking a business model to new technology and shifted consumer demands is not a simple task. It requires that you understand the context, see many links that need to connect for success to happen. This is a supercomputer task, something where the most simple human being (shout out to my deskmate 😉) outsmarts even the most advanced algorithm. To remind the reader, in this post I argued that: Executives should embrace the strategic advantages of superhuman A.I. technologies, knowing what can be automated and what not Companies should reposition there human workforce to focus on tasks that require supercomputer qualities Simple tasks that are well defined are being automated fast. So don’t be surprised when your competitor one day seems to have gotten superhuman powers, they probably did. Humans are still needed in your company. Let them focus on supercomputer tasks, let them be entrepreneurs. https://upscri.be/hackernoon/

Superhuman artificial intelligence versus supercomputer humans

At we make the fastest email experience in the world. We must therefore process massive amounts of text rapidly. We need to find links, validate emails, parse invitations, and much more. Superhuman , very Most programmers process text with . And rightly so: regular expressions are concise yet powerful. regular expressions But many programmers have also encountered the dark side of regular expressions. When regular expressions go wrong, they go devastatingly wrong. The Dark side of Regular Expressions Superhuman automatically converts email addresses into mailto: links. In this case, we defined an email address as any string that matches this regular expression: . You can read this as: “0 or more of any character except @, then an @ sign, then 0 or more of any character except @”. /([^@]*)@([^@]*)/ But while reading the of email addresses, we found some that did not match our regular expression. For example, if the username is quoted then it can also contain an @ sign. Consider . formal specification "joe@home"@example.com Now you might be thinking: “That’s crazy! Who would do that?!” But when you deal with user data, the unexpected happens all the time. Imagine you have 1 million users, and that each user has 50,000 emails. Between them, this is 50 billion emails. Even if it only happens once in every billion emails, that could still easily be 50 times! Wanting to do the right thing, we naïvely changed our regular expression to . This is exactly the same as before, but also allows the username to contain 0 or more quoted sections. /("[^"]*"|[^@])*@[^@]*/ It only took a few days before we received an email saying: “Please help! Superhuman is using 100% CPU and not responding…” We could see that the problem started after we changed this regular expression, and we could see that Superhuman broke on one email in particular. In some cases, the CPU would lock up for days. But why? It turns out that we had accidentally become vulnerable to (ReDoS). regular expression denial of service Regular Expression Denial of Service Theoretically, a regular expression is equivalent to a state machine that matches one character at a time. The state machine for our first email matcher looks like this: Example state machine for /[^@]*@[^@]*/ This state machine has 3 states: , , and . The machine starts in state . While here, the machine will match any character except and remain in state . If it encounters an sign, the machine will transition to state . A B $ A @ A @ B In state , the machine will match any character except , and remain in state . At the end of the string, the machine transitions to state . B @ B $ If the machine encountered an sign while in state , it would error as there are no matching transitions. The error would show that the string did not match the regular expression. @ B So what changed when we updated our regular expression? The state machine for our second email matcher looks like this: Example state machine for /("[^"]*"|[^@])*@[^@]*/ Do you see the problem? It’s not obvious, but we introduced non-determinism. In state , if the machine sees it has a choice: treat it as and transition to state , or treat it as and stay in state . A " " C any character except @ A There are some to this problem, but JavaScript and most modern programming languages take a dangerous approach: when the state machine has multiple paths, it will just choose one and continue. If that choice leads to the entire string matching, the machine will stop. If that choice does not lead to a match, it will and try the next path. theoretical solutions backtrack In the worst case, the state machine has to try every single possible combination of options before it can determine that there is no match. And the number of options very quickly becomes huge. In our example, every character doubles the number of possibilities. Our regular expression can therefore take (2 ) attempts to match, where is the length of the string. " O ⁿ n If you don’t believe me, open your browser console and type: let regex = /("[^"]*"|[^@])*@([^@]*)/ t = performance.now() regex.test('"""""""""""""""""""""""""""""""""""""""') console.log(performance.now() - t) // about 3 seconds. Each additional character doubles the time required. This 40 character string takes about 3 seconds on a high-end MacBook Pro. A similar string of just 64 characters will take more than 2 years — assuming you don’t run out of battery in the meantime! " In summary: If our regular expression is run against a valid email address, it will not backtrack and it will run very quickly. If our regular expression is run against the vast majority of common input, it will backtrack a little and it will still run quickly. If our regular expression is run against a very specific pattern, it will backtrack catastrophically and may never end. In other words, ReDoS manifests only in specific conditions, is catastrophic when it does, and — worst of all — cannot be caught by traditional testing as the conditions needed to trigger it are rare. The solution We turned to academia, and found that identifying ReDoS is still a relatively new area of research. One paper, however, was particularly useful. In “ ”, the authors describe a beautiful theoretical approach to this problem. They also provide a tool to test regular expressions. It works like this: Static Analysis for Regular Expression Exponential Runtime via Substructural Logics Compile the regular expression into a state machine. Look for ambiguity within the loops of the state machine execution graph. Run a bounded search of the execution graph to determine if these ambiguities can be triggered in a loop. If so, this would indicate a ReDoS vulnerability. A brilliant side effect of this strategy is that it generates an example string that will trigger ReDoS. This can be extremely useful for debugging, as you can see which parts of the regular expression are triggered. The original tool is written in OCaml and requires many dependencies to run. To make it easily available for everyday use, we created ! This is the OCaml tool wrapped in an HTTP API, and with some extra javascript-regex features. regex.rip With , you can create regexes with confidence. You can test your regexes so that users don’t have to. And best of all, you can avoid catastrophic downtime that you would otherwise only see in production. regex.rip Please feel free to use with all your regular expressions. If you want to contribute, please find the code on GitHub at . regex.rip superhuman/rxxr2 Thanks to Elena , Joy , Ruchi , Terin , and Islam for comments and suggestions on this post. At Superhuman we’re rebuilding the email experience for web & mobile. Think vim / Sublime for email: blazingly fast and visually gorgeous. If you enjoy working on user-facing problems that push existing solutions to their very limits — or beyond — join us! Learn more or email me .

How We Eliminated Regular Expression Denial of Service and How You Can Too

Will You Face Superhuman Opponents at the Next Olympics?

Disclosure: I am not associated with NeuraLink in any shape or form. As we’ve all have seen in the media and around the web, Elon has been having a rough few months. From sleeping on the factory floor in order to meet Tesla’s production objectives to consistently ticking off regulators and shareholders. It’s a wonder he has time for anything else in his life. But then a month ago, he showed up on the Joe Rogan Podcast. While talking about things like Bonobos’ mating habits and taking half a hit of a well deserved legal blunt, he told Joe that his Neurotechnology company, Neuralink, is going to be announcing some ‘interesting’ news in a few months: I think we’ll have something interesting to announce in a few months that’s at least an order of magnitude better than anything else. I think better than anyone thinks is possible. — Elon Musk What do we know so far about Neuralink? For starters, the companies ambitions are nothing short of the most Sci-Fi Sci-Fi movie. The company was founded in July 2016 by Musk and 8 others. Their short term goals are to treat serious brain diseases such as Alzeimers and Parkinsons and eventually get to enhancing the brain by “merging it with AI.” The co-founders of the company are all highly reputable individuals that each bring a very special set of cross-disciplinary skills to the table to create Neuralink. Their skills cover everything from biology to neuroscience to computer science. Here are the co-founders (other than Musk) involved in the company and a sample of their expertise: Ben Rapoport — Dr. Benjamin Rapoport graduated with an M.D./Ph.D. from Harvard Medical School. He also holds a Ph.D. in Electrical Engineering and Computer Science from MIT. Dongjin Seo — A neurotechnologist who has “created ultra-miniature, untethered, wireless neural implants (‘Neural Dust’) for brain-machine interfaces.” Max Hodak — According to his LinkedIn he “ studied biomedical engineering at Duke and worked in a lab studying multi-unit electrophysiology in rhesus monkeys to build brain-machine interfaces.” Paul Merolla — Has been the lead designer on more than 10 brain inspired chips including IBM’s TrueNorth chip. Philip Sabes — A researcher who through The Sabes Lab studies how the brain uses sensory feedback to maintain accurate and adaptive movement control. His lab also applies this research to the development of Brain Machine Interfaces (BMIs). Tim Gardner — A biologist who has worked on studying the neural activity in songbirds with the use of Brain Machine Interaces (BMIs). Tim Hanson — A Neurobiologist who taught himself about microfabrication methods and material science to develop some of the core technology at Neuralink. Vanessa Tolosa — A microfabrication expert, researcher of biocompatible materials and a Ph.d in chemical engineering. Together with Musk at CEO, this looks like the dream team of the brain enhancement race. What problem is NeuraLink trying to solve? The best way to understand the problem is to first understand how we’ve been progressing into an increasingly intelligent species. Elon says we‘re already cyborgs. The modern day human is able to answer almost any question you ask it within a matter of seconds. That futuristic device we all carry around connects us to an infinite pool of human knowledge. If someone makes a discovery, it can be shared with the human race within seconds on the internet. Combined with our devices holding access to this unlimited knowledge pool we‘re basically a giant inter-connected cybernetic organism composed of humans and machines. Terminator inspired cyborg. “your phone is already an extension of you, you’re already a cyborg.” — Elon Musk As our technology advances, our collective knowledge and intelligence only grows greater. A large part of this is due to us creating better communication systems to spread this knowledge across the network of humans. We’ve come a long way from our ancestors in terms of developing communication. The humans’ progression of communication looks roughly something like this: Started learning to make sounds. Sound waves. 2. Associated those sounds with the world around us. 3. Created language and passed down these sounds verbally as learning to be shared among others. 4. Drew symbols on surfaces to represent our world. A caveman chiseling a rock. 5. Created written language. Ancient Egyptian Hieroglyphics. Artfactory.com 6. Books. Available but scarce and accessible by few. One of the oldest surviving books in the world. A Jewish prayer book. Dated around 840 AD 7. Mass produced books. Accessible to many more. A large library of books. 8. Invented the telephone. Creation of another information sharing avenue. Telecommunicatons network. waitbuywhy.com 9. Invented Radio and television. Methods of communication using more senses. Early television. 10. Created Cell Phones — We’re no longer bound by physical location to communicate distantly. Early consumer cell phone. 11. The Internet — Another form of communication, possibly the most powerful of them all. A giant network of knowledge with almost everything we know as a species. A collective brain. visualization of the internet as a global brain with pathways of neurons and synapses. Opte Project We’ve accomplished a lot as a species by overcoming obstacles to create these clever forms of communication. By creating more of these mediums, we’ve been increasing the rate at which we’re able to exchange our knowledge with each other. Creating machines and the internet was one of the biggest leaps in this regard. That is where our next hurdle lies: Taking our knowledge exchange abilities with machines to the next level. Although we’re already cyborgs, the rate at which we can communicate with our machine counterparts isn’t very high. Musk says that we’re facing a bandwidth issue here. “the communication rate between you and the cybernetic extension of your self, that is your phone, your computer is slow, it’s very slow.” — Elon Musk Now what if we could eliminate all the slow mechanical aspects of accessing the information available to us through machines? For example, instead of typing something out with our hands, having the machine execute our request, then reading the information retrieved with our eyes, what if we could just think about it and access it instantly from the machine? What if our minds could have access to a data storage device and instantly query it with our thoughts? With these abilities we would be able to learn and remember exponentially better than using our existing physical methods of communicating with these interfaces. Not only that, it could help countless amounts of people who are suffering from brain related illnesses such as Alzheimers, where your memory is usually the first to go. This bandwidth issue is one of the problems NeuraLink is hoping to tackle in the long run. One solution they’ve devised is to use a Brain Machine Interface (BMI) which will link from your cortex to a symbiotic AI to drastically increase your cognitive capabilities. Think Limitless but instead of a pill to juice up your brain, it’s an AI chip implanted somewhere in your body. An AI extension of yourself as Musk puts it. Ok, the Limitless example might be a stretch but I’m sure we’ll live that future some day. When asked about the long-term goals of NeuraLink’s technology, Musk had this to say: “It will enable anyone who wants to have superhuman cognition, anyone who wants. This is not a matter of earning power, earning power would be vastly greater after you do it. Anyone who wants can effectively do it, in theory.” — Elon Musk What is a Brain Machine Interface (BMI) and what exactly does it do? Let’s first take a quick look at how our brain works and how scientists currently interact with it. The average brain is composed of 100 billion neurons which form a information highway. A neuron is made up of 4 main parts: The Soma (Cell Body), Dendrites, Axons and Axon Terminals. In the diagram below of the single neuron, the Soma contains Dendrites which is where the neuron listens and receives signals from other neurons. There could be hundreds of these for each neuron. The received signal then triggers what’s known as Action Potential within the neuron which is like electrocuting it. As our neuron is electrocuted, it sends out more signals ahead to other neurons via it’s Axon Terminals which are making contact with other neurons’ dendrites thereby electrocuting them. The points at which the neurons connect are known as synapses. A standard pyramidial neuron. Neurons come in many different types. This system of communication operates similar to a computer’s binary language of 0s and 1s. The neuron is either excited with Action Potential or it’s not. The difference is that our brain constantly changes it’s neural connections allowing some neurons to be better connected to the rest of the network than others and vice versa. This change in connections is what we know as learning. When you learn something new, the neural network in your brain is creating action potential in neurons that never got much love before and in some cases generating brand new neurons. Neurons experiencing Action Potential. Now that we have an idea of how the neural network in our brain communicates, we can start understanding how to record and decipher it. Interacting with the brain involves 2 categories: 1. Recording information we want from the brain. 2 . Inputting the information we want into the brain. Currently, scientists and healthcare professionals are using tools like EEGs and fMRIs to analyze brain activity. These tools can provide great clues as to what’s going on in there but they are limited in the detail of data they can collect. The human skull is part of the problem in that it blocks much of the electrical signals we’re able to record. To really understand what it’s doing scientists need to get more detailed access to these signals. Multielectrode Array which is implanted into beneath the skull. One way to do this is to implant a multielectrode array underneath the skull. Sounds scary as hell, but it’s already been done. In the image above, the tiny needles on the electrode record the signals generated by the brain and are sent off to be deciphered by the machine interface. Size comparison of a multielectrode array. This is where we currently are with this technology and it’s quite amazing. It’s already shown the potential to change many lives. Watch the video below to see for yourself. Knowing where we are with this technology now gives us a better idea of what we could expect from this upcoming NeuraLink announcement. Now let’s add AI into the mix and talk about why Musk has been warning of “The Singularity” on almost any public forum he shows up on. What is the singularity and why should we be worried? T-800 Image by Dick Thomas Johnson According to Wikipedia: The (also, simply, ) is the hypothesis that the invention of artificial superintelligence (ASI) will abruptly trigger runaway technological growth, resulting in unfathomable changes to human civilization. technological singularity the singularity — Wikipedia Above, we talked about how humans are essentially cyborgs because we’re part of this giant ever expanding cybernetic organism made up of humans and machines. Over time, we’ve been able to increase our communication bandwidth which has exponentially increased our abilities as a human collective. We’ve come to the point where we are now creating a digital replica of our own cognitive abilities, creating machines that can “think” and do like us. So how far are we humans in this process of creating the artificial us? To compare, here are the 3 major parts of our brain that exist in us and what they do: Three parts of the brain. (clinical-partners.co.uk) Reptilian Brain — The oldest part, controls vital functions like breathing, heart rate, body temperature. Limbic System — The mammalian part of the brain which is used for learning, memory, emotions and behavioral responses. Neocortex — The “thinking” part of the brain which is involved in conscious thought, reasoning, language, motor commands and sensory perception. If we were to try to measure how close we are, some might say that we’re in the neighborhood of the Limbic System. Artificial neural networks could be a good example of this. Their job is to take in data and learn from it as good as or even better than a human. And that’s the part where it gets scary. These AI networks have instant access to what ever information we give it. When it comes to bandwidth, machines have exponentially higher than us limited only by their processing power. Imagine a machine making a discovery and sharing it pretty much instantly with the entire network of other machines. Every node in the network would be equally intelligent within seconds. That makes you wonder… If we’ve come so far in such little time with our limited learning abilities, how much could a sophisticated AI network accomplish if given access to the wealth of human knowledge that already exists? I find this question both full of great hope for our future and at the same time I’m thinking SkyNet. This is why Elon Musk said: “If you can’t beat it, join it” -Elon Musk When I started thinking about this, I realized, Musk has a point. He says once we’re able to successfully merge with our symbiotic AI counterparts, “the outcome for humanity will be the sum of human will.” — Elon Musk I like the sound of that. This article was inspired by Wait But Why’s extremely in-depth post about Neuralink, click here to check it out. I highly recommend it. Click here for Joe Rogan’s Podcast with Elon Musk where you can watch him take a blunt hit and talk about all this.

Elon Musk wants to give you super human cognitive powers

Photo by Andy Kelly on Unsplash Everyone is saying AI is coming for our jobs, and it’s coming fast. Will AI replace humans and create an inevitable worldwide unemployment crisis? Humans are good at surviving and excelling in the real world . . . aren’t they? When hunting and gathering were our only requirements, the development of agriculture didn’t mean the end of life as we knew it. We adapted and evolved. The industrial revolution didn’t result in unemployment armageddon. Instead, humans created more jobs and reinvented themselves with the help of new technologies. The advent of the Internet and the knowledge economy didn’t permanently put us out of work, either. Instead, each of these innovations made us more productive with elimination of very few job categories. Technology and innovation makes our lives easier and helps us do our jobs better. That’s what AI has in store for us. It makes us more efficient — in effect, superhuman. #1. AI frees us from repetitive tasks There are too many repetitive tasks in our jobs. For example, take recruitment. A recruiter’s job is to find talent for organizations. This role involves personal interaction so the recruiter can build relationships with clients and candidates. What they spend way too much time doing is sorting through resumes (often hundreds or thousands), searching through job sites, scheduling interviews, making phone calls, and sending emails. . That frees up the recruiter to take care of what’s really important, like developing relationships with top candidates and getting them into the best jobs. AI can easily handle repetitive recruiting tasks, and it can highlight the best candidates — those with the greatest chance of success in a particular organization That’s win-win situation. #2. AI isn’t smart enough to do creative work AI can plow through a million psychology textbooks in a fraction of a second, then tell you all symptoms of and treatments for depression. But only a human can read a face and instantly know the right thing to say. This is emotional intelligence, which AI simply doesn’t do. And it never will. Our nimbleness to react instantly in very creative ways is unmatched by any computer. The auto suggest feature on your smartphone uses AI to predict the next word you want to type. It helps you type your message faster, but it can’t fully take over and chat with a colleague by feeling, understanding and appropriately reacting to human emotions. Even Google Duplex that has got people scared is designed to operate in very specific use cases such as hair salon booking, restaurant reservations, and holiday hours. AI will only work — and learn — within a given set of constrained parameters, unless is achieved. singularity Google’s AI search algorithm has transformed the research process. You no longer need to crawl through dusty libraries or know the dewey decimal system. Google crawlers and algorithms deliver what you need with uncanny accuracy. You get all the best query results through Google’s ranking system, without any delays for thought or clumsy boolean query strings. AI makes us more effective and more efficient. It gives us back the time we need to do creative work, including reviewing our search results and assimilating the information. #3. AI allows us to be better at our jobs by making us more human Take sales, for instance. Selling is an art that takes incredible finesse to do well. It’s the marketer’s most basic directive and it’s an inherently human function. Salespeople waste time qualifying prospects, sending follow-up emails, and updating client relationship management (CRM) software instead of being dedicated to human interaction and rapport-building. Those with a natural knack for selling might be able to keep up, but the rest of your team isn’t going to be as productive as they should be. When you add AI to the equation, the job gets easier and your sales force gets stronger because now they have time to study and learn from your high performers. AI takes care of unproductive overhead so salespeople can focus on essential tasks like listening and asking questions, empathizing with prospects, and finding creative solutions to problems. When we’re freed from menial work, the world slows down around us. We have more time to strategize, relate with clients, and optimize our processes to increase productivity and get a greater return on investment. #4. AI creates new industries and jobs Who could have predicted that a breed of converted Facebook junkies called social media marketers would create a $31 billion industry leveraging social media? Social media jobs were non-existent 10 years ago. Today, social media managers are earning six figures, on par with the best technology workers. Let me tell you a story from the eighteenth century, when the Spinning Jenny was invented. Workers in the spinning profession were terrified of losing their jobs. The Spinning Jenny reduced the amount of work required to spin thread, which led to a fall in the price of textiles. The cotton spinning workforce was so afraid of losing their jobs that they . Today we have cheap, mass-produced cloth because of the foundation that was laid by the Spinning Jenny — and, contrary to what the workers feared, the affordable thread resulted in uses that expanded the industry even more. broke into the inventor’s house and smashed his machines Considering the improved lifestyle we’ve experienced since this technological disruption, shouldn’t we be similarly excited about AI? Expanding productivity will cause industry to flourish in ways we can’t even imagine. Here’s what’s AI is already doing: AI software is improving elder care as demographic trends overwhelm the care industry. AI-enabled robots are taking over menial and monotonous tasks. It’s leading to more efficient money management, applying the best investment strategies. It’s serving and protecting citizens by enabling effective and unbiased law enforcement. New devices and new technologies will give birth to new industries and jobs, some that are unfathomable today. #5. AI can’t solve healthcare’s major problems You’ve probably seen the Black Panther movie, where Vibranium powered AI solutions are so common that a gunshot to the spine takes only a day to heal. So you might think AI will provide a cure for AIDS next year. Well, that’s just fantasy. There are highly intelligent AI systems like IBM’s that can memorize a bazillion ailments and recommend treatments. But it will be decades before AI finds a cure for cancer or old age, because it lacks the creativity and spontaneity that sparks human genius. Watson We won’t have instant discoveries in healthcare or major world problems just because of AI. Trials, experiments, peer review, and many million lines of code stand in the way. That’s all done by humans. And progress will be slow. But what AI will do is accelerate the process significantly. With artificial intelligence, extremely detailed models and simulations of human cells and other physiological components are possible. These models will boost experiments to the next level and save years of research. We’ll be able to reduce death rates and prolong human life. AI will help us work faster, eliminate impractical solutions, and narrow the pool of promising alternatives. Artificial intelligence: what does the future hold? One of the most interesting things about AI is that, unlike static code, it is programmed to learn. That means it can evolve all by itself. A lot of commentators predict that sometime in the very near , AI will achieve singularity. future At the stage of singularity, AI will become smarter than humans. It might even be able to think, essentially evolving into a new species. If this fiction steps out of sci-fi movies and becomes reality, the human race will face a challenge. Ethicists constantly debate what to do about it. None of that changes the benefits of the explosive AI innovation going on in Silicon Valley and across the globe. IBM’s Watson has tried its hand at several careers, including professional sports scout and medical diagnostician. But, most famously, Watson is a Jeopardy! winner and chess champion. Each of these scenarios was a narrow construction that allowed Watson’s brute force computing power to demonstrate expertise. There’s even AI tech that handles complex problems in financial services. One Hong Kong company . AI bots have also made an appearance in the literary industry, creating interesting work in and . appointed an algorithm as its director poetry prose Not to fear — we can sit back comfortably and look forward to a future where humans and AI work together to solve the world’s problems in real time. Artificial intelligence: legal implications and IP protection AI is a disruptive innovation and isn’t limited to the tech and business sectors. Businesses are getting savvy to the fact that a lot of tricky situations are ahead, especially concerning the legal implications of artificial intelligence. AI pioneers need to protect their intellectual property. Regulation has become a big issue, especially after incidents like the recent fatal accident in Arizona involving an self-driving Uber car. The laws revolve around driver fault, but how do you prove fault when no human is driving the car? Lawmakers are already working to push through legislation to regulate self-driving vehicles, while at the same time trying not to squelch innovation. Issues with IP protection are especially close to home for CEOs and senior executives of companies that design and develop artificial intelligence systems. Today’s tech race is competitive, and intellectual property is its most valuable commodity. The smartphone industry, for example, has shown us that the interval between innovation discovery and its appropriation by competitors can be as short as a few months, in part because so many AI developments are open source. — with AI development getting so much attention. — Businesses need to have a patent strategy to protect their innovations if they want to maintain market dominance or evolve to become a dominant player in their markets Executives can no longer afford to scrape by with minimal knowledge of IP protection while disruptive is continually developed and protected. Focused IP protection strategies need to be unboxed and implemented even before an innovation is market-ready. It’s the only way to stay ahead of the wave AI investment. technology Patents, copyrights and trademarks offer legal protection. But it can be difficult to determine which type provides the best protection for each AI businesses investment. The big question is: How is your company differentiating yourself with your technology? Is your innovation new source code, improved technical functionality, a strong brand, or something completely different? Startups often leverage open-source software (OSS), without the burden of legacy systems, in order to quickly develop AI innovations. OSS licenses differ in their implications on IP protection strategies. Some licenses let you protect your innovation with patents, but restrict you from enforcing them. Most prevent you from excluding other startups or established companies from using the source code you develop on top of the OSS. Even after you’ve decided on a form of protection, seeing it through can present serious challenges. Obtaining a patent in the US typically takes anywhere from 3 to 5 years. While the patent will protect the innovation for up to 20 years, any advantage the company has may be lost while they’re waiting for the patent to go through. Three years is an eternity in Silicon Valley! That’s where fast-tracking your patent can speed the patent office’s decision to a year or less. The process for obtaining a patent can be confusing and can potentially take years to complete. A landmark decision by the US Supreme Court in 2014 puts a big question mark on the possibility that AI solutions can be patented. That’s because the court doesn’t regard ‘abstract ideas’ like ‘disembodied algorithms and formulas’ as patent-eligible. In addition, the very real question of ownership of original works will emerge in the near future when AI helps with innovation. A photographer by animal rights activists who claimed that the copyright in a selfie taken on his camera by a monkey belongs to the monkey, not to him. The future patent tussle won’t be between humans and monkeys, but between humans and AI. was recently sued Current AI practices teach bots to evolve through machine learning. This means they can evolve to the point that they’ll produce output radically different from their source code and, in some cases, even rewrite their own code. When that happens, who will own the copyright, the code writer or the bot? What if the improvement is patentable? The best way to tackle these legal and IP issues is to . An expert can help you develop a robust strategy for protecting your company’s intellectual property. approach a legal strategist early in the innovation process Businesses, including startups, can’t afford to delay any longer. The first to file for a patent will own that innovation. Trends predict that dominance in AI will be determined not by who can innovate fastest, but who can protect their IP first. Conclusion Artificial intelligence may eliminate some jobs but it will also create new ones, and work that is done by humans will be more thoughtful, rewarding and fulfilling. AI will make us beyond human — in effect, superhuman. We’ll march on and evolve because that’s what we do. And we’re good at it.

f5 Reasons AI Won’t Replace Humans… It Will Make Us Superhuman

Christopher Columbus, Conquistadors and Genetic Engineering Historically humanity has been driven by one primary motivation: survival. Today we have evolved past scarcity and live in a world where those living in “first world” countries do not have to worry about starving to death. If anything overconsumption kills vastly more than does hunger. Once our survival needs are met, we as a species think bigger: money, power, maybe even self-actualization. Unfortunately many stay stuck in the first two, failing to achieve a higher level of being. While historically this was okay, it is becoming increasingly less so in the modern era. Our power exceeds our grasp… NOTE: this is not a doomsday post about the many threats humanity faces. This is more important. Let me explain. A history of inequality You could argue we are living in the most equal era of human history. Conditions around the world are vastly improved over just 10, 20, 50 years ago. And compared with much of human history, we are living in near utopia — a safer, healthier world than any of our ancient ancestors could have ever imagined. Despite these advancements, we may be entering a brave new world built on slavery — but first we need a little perspective. Slavery sells While slavery is morally indefensible, historically it has been incredibly popular and profitable. And surprise surprise, it seems to have cropped up throughout human history again and again ( ). Source, and shocking timeline Athens banned debt based slavery in the early 6th century BC. That was 2500+ years ago. And by 12AD, slavery was banned throughout the majority of Greece and the Chinese and Roman empires. But like all evils with money to be made, it returned with a vengeance. Then again during the Middle Ages kingdoms throughout Europe needed to ban slavery again, including the entire Christian “kingdom.” Every time humanity discovered new lands, new rules needed to be written. The new “savages” had little to no protection and thus slavery began anew — again and again and again — throughout the colonial era, the modern era and through to today. Finally in 2008, Kazakhstan, the last of the major countries signed the 1926 Slavery Convention accords, a full 82 years after being written… It took a Civil War to end slavery in America. ( ) Factory farming is another great example of exploitive “slavery”. Rather than elaborating further, see the following interview with Mike Selden of Finless Foods, a company synthetically manufacturing fish meat It seems like market forces alone aren’t always enough to stop greed. Exponential evolution The advent of CRISPR will likely be THE MOST IMPORTANT technology of our era, vastly outstripping mere communication tools like the internet and mobile phone. technology If there is a god, he is probably coding on CRISPR-esque computer, shaping the world with lines of genetic code. Now our power really exceeds our grasp… And whether the UN and nations of the world decide to ban testing or not, it will be virtually impossible to prevent human experiments. The prospects of immortality, longevity, sex, skinny waists, super intelligence and much more will drive us to self-experiment. And that is to say nothing of curing cancer or saving a loved one. genetic The cat is out of the bag, the question is how long it stays still. Innovation is expensive When computers first came out, none but corporations and governments could afford the ~$3M price tag. It was only after several decades of experimentation, improvement and cost reduction that personal computers began to enter the home. The tech was too big and too expensive for the masses. That was fine. And while this created inherent imbalances between those with access to computers (like Bill Gates) and those without access, the disadvantage was clearly not insurmountable (see Google, Amazon, Facebook etc…). But what if this same imbalance plays out with access to genetic engineering. Evolution typically happens on the order of thousands or hundreds of thousands of years, but we are entering an era when it could be at most a generation, arguably much less, in order to see enhancements. If only the super-rich can afford edits and enhancements (either for themselves or children), what happens to the playing field. Increase intelligence, strength and lifespan enough and we are dealing with a different species entirely. Humans are only ~1.2–1.5% different genetically than chimps, yet we keep them in cages… Every time humanity discovered new lands, new rules needed to be written. The new “savages” had little to no protection and thus slavery began anew — again and again and again — throughout the colonial era, the modern era and through to today. This is a distinct possibility if humanity continues to play in a capitalistic world focused on scarcity. If enhanced humans don’t have enhanced emotional and spiritual capacity, we may be looking at an accelerating era of inequality unlike any we have known before. Talk about the 1%. And if they never die, the inequalities grow exponentially from financial successes, improved technologies and zero estate/death taxes… These same differences will likely manifest themselves in mechanical human augmentation as well, creating another source of us vs them mentality so popular throughout human history. NOTE: Technology might save us (sort of) The thing is, if we are smart enough to succeed with enhancing ourselves, we are probably smart enough to succeed in creating robots capable of most if not all basic tasks/labor. This could lead to an era of abundance unlike any other. But it could also be our undoing — not in terms of an AI takeover (possible but not relevant to this conversation) but by destroying our humanity. We do not understand consciousness, the hard problem eludes us. If humanity successfully creates robots (which at this point seems inevitable), at what point (if any) will they become sentient? Would trading miserable human slaves for miserable metal slaves be any more moral? While it is entirely possible we may never create conscious robots, we will never know. We cannot perceive consciousness, only assume whether or not it exists. If we believe we have created conscious beings, can we in good faith keep robots as unpaid, ever working slaves (and sex robots)? Either we erode our humanity or our economy — could society suddenly eliminate automation and still survive? What about in 20 years? Closing thoughts We are entering into the most interesting era of human existence where our powers and abilities as individuals become exponential amplified — whether for good or for evil. By considering societal and ethical issues before they arise, we can attempt to control the outcomes and align interests for the betterment of humanity. With CRISPR and AI in particular though this is hard. Both are winner-take-all technologies where all that matters is getting there/succeeding first. I foresee gross human rights violations as experimental CRISPR edits are illegally tested and perfected — similar to what pharma companies do to 3rd world “volunteers” but with more dangerous and promising “drugs.” “The times they are a-changin’ “— Bob Dylan It is these topics and more we tackle on , the podcast exploring the edges of human understanding. If you are interested in learning more about the converging exponential technologies of the future like genetic engineering, space travel, AI, human longevity and more, be sure to subscribe on or on FringeFM iTunes , Stitcher our site. Learned something? Click the 👏 to say “thanks!” and help others find this article. Hold down the clap button if you liked the content! It helps me gain exposure . Clap 50 times!

Superhuman Slavery

From OpenAI to Closed AI: Custom Chips Are Closing The Doors—What’s Next? 

Ад OpenAI да закрытага AI: Карыстальніцкія мікрасхемы зачыняюць дзверы - што далей? 

De l'OpenAI a la IA tancada: els xips personalitzats estan tancant les portes: què passa? 

Von OpenAI zu Closed AI: Kundenspezifische Chips schließen die Türen – was kommt als Nächstes?

Από το OpenAI στο Closed AI: Τα προσαρμοσμένα τσιπ κλείνουν τις πόρτες—Τι ακολουθεί; 

De la IA abierta a la IA cerrada: los chips personalizados están cerrando las puertas. ¿Qué sigue?

Az OpenAI-tól a zárt mesterséges intelligenciaig: Az egyedi chipek bezárják az ajtókat – mi a következő lépés? 

Dari OpenAI ke AI Tertutup: Chip Kustom Menutup Pintu—Apa Selanjutnya?

OpenAI からクローズド AI へ: カスタム チップが扉を閉ざす — 次は何が起こるのか?

Kobanda na OpenAI tii na AI oyo ekangami: Ba puces personnalisées ezali kokanga ba portes —What’s Next? 

له OpenAI څخه تړل شوي AI ته: دودیز چپس دروازې بندوي — بیا څه دي؟ 

От открытого ИИ к закрытому ИИ: кастомные чипы закрывают двери — что дальше?

Від відкритого штучного інтелекту до закритого штучного інтелекту: спеціальні мікросхеми закривають двері — що далі? 

Winter(Physics) is Coming It now looks like Large Language Models running on the GPT technology may be nearing a plateau in their development. This is from popular Physicist Dr. Sabine Hossenfelder’s latest YouTube video. https://www.youtube.com/watch?v=AqwSZEQkknU&embedable=true Dr. Sabine cites a couple of sources, including Bill Gates, who said last year that GPT 5 won’t be much better than GPT 4. And Prof. Gary Marcus says he has been drumming this particular beat for years. This goes against all the Ray Kurzweilian hype by Sam Altman and Ilya Sutskever as seen in the video. But if merely beautiful mathematics won’t spur the progress of Physics, as Dr. Sabine always reminds us, then merely wishful group-thinking won’t give us a superhuman artificial general intelligence (AGI). Why isn’t throwing as much data, computing power, and human supervision as can be afforded, going to birth us a superhuman AGI? Dr. Sabine has a ready answer. The problem, she explains, is the Decoupling of Scales. The Decoupling of Scales The world is broken up into scales. In the physical world, there is the scale of land animals like us humans. Then below that we have microbes, then molecules, atoms, and subatomic particles. In the opposite direction, we have aircraft and skyscrapers, mountains, planets, stars, solar systems, galaxies, etc. In the world of data, we have high-resolution data e.g. categories like “objects that can be touched”, “animals”, “cats”. Then low-resolution data to further expand the above categories e.g. “cat species”, “cat patterns”, “cat behavior at 8:00 am on Monday,” etc. There are other taxonomies, however. Some data is online and some is offline in books, papers, and people’s minds. We may have all the knowledge we desire at the high-resolution scale of human language creation on the internet. But that doesn’t equate to the much bigger, low-resolution data source that is human language creation in the world. Thus, Sam Altman’s AI is like Aristotle in the image below. Hoping to discover the Standard Model of Physics by looking at his hand. But without something like the Large Hadron Collider to poke subatomic particles and see the strange new data hiding within, Aristotle is at a loss on how to proceed. A Large Hadron Collider for AI GPT architecture, unveiled to the world only 2 years ago with the phenomenal ChatGPT, is likely nearing the end of its “improvement-ness”. Perhaps it is time to seriously consider passing the baton, and the VC funding, to something else. I am rooting for thermodynamic computing, currently being explored by companies like Extropic, which works by leveraging thermal noise to create random samples of data as in Stable Diffusion technology, then denoising using statistical analysis. In fact, in any neural network, the initial weights are usually initialized with random noise. During training, backpropagation shapes the final weights from this random noise - writes Laszlo Fazekas. The advantage lies in the chips operating at scales smaller than those of traditional digital computing, where quantum effects result in abundant thermal noise. Thermal noise has always been plenty, anyway. My kitchen stove is full of thermal noise. The real problem has always been how to understand it better through statistical chunking of its data, without throwing away data during regression analysis. My guess? Some sort of massive cloud computational AI training leveraging smartphones, bitcoin nodes, TV sets, radios, name it. Currently, allowing the AI some “thinking time'“ so that it runs longer inferences on the same data, is how AI companies are squeezing out some more juice. For example with ChatGPT’s o1 model. But maybe even training time should be extended. Maybe to get great AI, like a child, it needs decades of training from data collected from all over. Collected and slowly siphoned into a condensed AI model in a few servers. Sam Altman might need those trillions of government dollars for such a plan. Alas, they will likely not be easy to get. The D.O.G.E might not bankroll him (if xAI is suffering the same troubles). Meanwhile, Bitcoin is still growing As the above graphic shows, very few people are using Bitcoin. Yet it is already at $90k. Data from Bitbo shows that in fact, only 400,000 people are using Bitcoin daily. And only 106 million own Bitcoin. Contrast this with the 314 million people who have used AI tools as per data from Statista. ChatGPT alone has millions of users per day. Probably more than the entire Bitcoin blockchain combined! What’s more? AI amassed the bulk of these users in less than a year. This feels wrong. Bitcoin took years to grow. People sweated, lost, DCA’d, had faith. Maybe that’s it. ChatGPT has grown too fast. Now it has slowed down and everybody is sour. Sorry, ya’ll, there is still work to do. AGI will not be so easy to grasp, and GPT scientists should learn to hodl like Bitcoiners. Winter will come and go, but right now, it is time to pay the piper. Winter(Physics) is Coming It now looks like Large Language Models running on the GPT technology may be nearing a plateau in their development. This is from popular Physicist Dr. Sabine Hossenfelder’s latest YouTube video. https://www.youtube.com/watch?v=AqwSZEQkknU&embedable=true https://www.youtube.com/watch?v=AqwSZEQkknU&embedable=true Dr. Sabine cites a couple of sources, including Bill Gates, who said last year that GPT 5 won’t be much better than GPT 4. And Prof. Gary Marcus says he has been drumming this particular beat for years. This goes against all the Ray Kurzweilian hype by Sam Altman and Ilya Sutskever as seen in the video. said last year says But if merely beautiful mathematics won’t spur the progress of Physics, as Dr. Sabine always reminds us, then merely wishful group-thinking won’t give us a superhuman artificial general intelligence (AGI). Why isn’t throwing as much data, computing power, and human supervision as can be afforded, going to birth us a superhuman AGI? Dr. Sabine has a ready answer. The problem, she explains, is the Decoupling of Scales . Decoupling of Scales The Decoupling of Scales The world is broken up into scales. In the physical world, there is the scale of land animals like us humans. Then below that we have microbes, then molecules, atoms, and subatomic particles. In the opposite direction, we have aircraft and skyscrapers, mountains, planets, stars, solar systems, galaxies, etc. In the world of data, we have high-resolution data e.g. categories like “objects that can be touched”, “animals”, “cats”. Then low-resolution data to further expand the above categories e.g. “cat species”, “cat patterns”, “cat behavior at 8:00 am on Monday,” etc. There are other taxonomies, however. Some data is online and some is offline in books, papers, and people’s minds. We may have all the knowledge we desire at the high-resolution scale of human language creation on the internet . But that doesn’t equate to the much bigger, low-resolution data source that is human language creation in the world . human language creation on the internet human language creation in the world Thus, Sam Altman’s AI is like Aristotle in the image below. Hoping to discover the Standard Model of Physics by looking at his hand. But without something like the Large Hadron Collider to poke subatomic particles and see the strange new data hiding within, Aristotle is at a loss on how to proceed. A Large Hadron Collider for AI GPT architecture, unveiled to the world only 2 years ago with the phenomenal ChatGPT, is likely nearing the end of its “improvement-ness”. Perhaps it is time to seriously consider passing the baton, and the VC funding, to something else. I am rooting for thermodynamic computing , currently being explored by companies like Extropic, which works by leveraging thermal noise to create random samples of data as in Stable Diffusion technology, then denoising using statistical analysis. thermodynamic computing In fact, in any neural network, the initial weights are usually initialized with random noise. During training, backpropagation shapes the final weights from this random noise - writes Laszlo Fazekas. in any neural network , the initial weights are usually initialized with random noise. During training, backpropagation shapes the final weights from this random noise - neural network backpropagation The advantage lies in the chips operating at scales smaller than those of traditional digital computing, where quantum effects result in abundant thermal noise. Thermal noise has always been plenty, anyway. My kitchen stove is full of thermal noise. The real problem has always been how to understand it better through statistical chunking of its data, without throwing away data during regression analysis. My guess? Some sort of massive cloud computational AI training leveraging smartphones, bitcoin nodes, TV sets, radios, name it. Currently, allowing the AI some “thinking time'“ so that it runs longer inferences on the same data, is how AI companies are squeezing out some more juice. For example with ChatGPT’s o1 model. But maybe even training time should be extended. Maybe to get great AI, like a child, it needs decades of training from data collected from all over. Collected and slowly siphoned into a condensed AI model in a few servers. Sam Altman might need those trillions of government dollars for such a plan. Alas, they will likely not be easy to get. The D.O.G.E might not bankroll him (if xAI is suffering the same troubles). Meanwhile, Bitcoin is still growing As the above graphic shows, very few people are using Bitcoin. Yet it is already at $90k . already at $90k Data from Bitbo shows that in fact, only 400,000 people are using Bitcoin daily. And only 106 million own Bitcoin. Bitbo Contrast this with the 314 million people who have used AI tools as per data from Statista . ChatGPT alone has millions of users per day . Probably more than the entire Bitcoin blockchain combined! data from Statista has millions of users per day What’s more? AI amassed the bulk of these users in less than a year. This feels wrong. Bitcoin took years to grow. People sweated, lost, DCA’d, had faith. Maybe that’s it. ChatGPT has grown too fast. Now it has slowed down and everybody is sour. Sorry, ya’ll, there is still work to do. AGI will not be so easy to grasp, and GPT scientists should learn to hodl like Bitcoiners. Winter will come and go, but right now, it is time to pay the piper.

The AI Winter is Near, Courtesy of Physics

Der KI-Winter steht vor der Tür, mit freundlicher Genehmigung der Physik

El invierno de la IA se acerca, cortesía de la física

به لطف فیزیک، زمستان هوش مصنوعی نزدیک است

Tekoälytalvi on lähellä, fysiikan luvalla

एआई शीतकाल निकट है, भौतिकी के सौजन्य से

物理学のおかげでAIの冬が近づいている

AI қысы жақын, физиканың рұқсатымен

Зима искусственного интеллекта уже близко. Предоставлено Physics

Iyo AI Chando Yava Pedyo, Nekuda kweFizikisi

Dimri i AI është afër, me mirësjellje të fizikës

AI குளிர்காலம் அருகில் உள்ளது, இயற்பியல் உபயம்

AI qishi yaqin, fizika izni bilan

I recently Googled And IBM.com gave the top result. As it explains, smart contracts automate coded actions on a cryptocurrency platform (blockchain). So I’m thinking trader bots and escrow agreements. The result also says they can automate a workflow. So hourly, weekly, or monthly (cryptocurrency) payments to a payroll. OK. I guess there is something smart about smart contracts. Coz people usually handle the aforementioned tasks. Then I got to thinking, can smart contracts be as smart as artificial intelligence algorithms? Yes, I Googled. It’s not a thing. Every site talks of a smart contract combined with AI. I want it to become AI. I know, we all want things. If wishes were horses. Nonetheless, I have a few quirky similarities to share. Starting with the cover image. Notice how the Rootstock logo looks a lot like an Artificial Neural Network? I did. Bitcoin’s supply mechanics and the Boltzmann distribution Rootstock is a Bitcoin side chain which means at its heart, it is Bitcoin’s supply mechanics that govern the halving and the difficulty adjustment and more. The true engine of all its smart contracting. While Rootstock was chugging along its Bitcoin rails, on the other side of town, the Nobel Prize in Physics was being awarded to John Hopfield and Geoffrey Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks.”- https://www.nobelprize.org/. Their work in machine learning led to GPTs (Generative Pre-trained Transformers), a now famous architecture in major AI systems today. What’s interesting is that foundational to all Artificial Neural Networks (ANNs), as mentioned in the Advanced information section of NobelPrize.org, is something called the Boltzmann Distribution. Below is its mathematical equation to capture the energy E of an ANN; Notice the similarity to the Bitcoin supply equation below (with 3 parameters too! Replace W_ij above with 210,000, S_j with 50, and S_i with 1/2^i) ; What can I say, there is something Bitcoin-like to AI. Both the Boltzmann distribution and the Bitcoin supply equation model an exponential decay/reduction. The former is generalized for the energy of a closed system with n particles, while the latter models the number of bitcoins available to be mined. This too is a closed system with n = number of halvings. The particles can be as many as they want to be (so whether Bitcoin is a planet-sized computational network or a modest network of a dozen PCs, doesn’t matter). But they will all decay in 32 halvings to mine all the bitcoins that can possibly be mined. Approximately 21 million. Now bitcoins are mined in a system that unlike standard computer systems, where linguistic logic coded by a programmer dictates the state of 0s and 1s, this machine changes when it is fed energy... - Sydney Bright, BitcoinMagazine. How smart are Rootstock smart contracts? We have seen that both Rootstock smart contracts and AI algorithms consume energy. So the question is, how well do they consume the energy? Well, both consume it pretty well to make some good money. They are both smart money algorithms in that regard. To see an instance of this, let’s put Rootstock next to ChatGPT in a certain way. Look at the below statistics. Of course, ChatGPT has more users, more money flowing in, and more employees worldwide. It is smart. We all know this. But pound for computational pound, Rootstock does hold its own. Let’s make some assumptions like a Physicist. For example, we can see that Rootstock.io has 261,406 monthly transactions. Assume they are from 261,406 different people. Meanwhile, ChatGPT has 800 million monthly users, generating billions of prompts (similar to transactions) every month. Let’s take their number of monthly users as equivalent to their monthly transactions. Now let’s divide market cap by these monthly figures (this is a made-up metric, but a good one): Rootstock’s “Market cap per monthly user” = $380,000,000 / 261,406 = $1453 ChatGPT’s “Market cap per monthly user” = $157,000,000,000 / 800,000,000 = $196 As you can see, ChatGPT is not doing better than Rootstock in this calculation! And if those 2740 RBTC pegged to BTC soon climb to $100k+ per RBTC (courtesy of the Trump Administration), you bet, ChatGPT will learn all about it. And OpenAI just might consider allying with a newly-unicorn Rootstock to provide paid ChatGPT services in RBTC. Bottom line similarity: Both Rootstock smart contracts and ChatGPT are smart money algorithms with good performance metrics. The former caters to a smaller clientele (Bitcoiners) with deep pockets, while the latter caters to a larger clientele (the world’s literate population) with relatively modest pockets. Speedsters In the Rootstock white paper, while the average confirmation time of BTC transactions is 10 minutes, Rootstock’s RBTC, which holds a 1:1 on-chain peg to BTC, takes only 30 seconds in average confirmation time. This implies that RBTC transactions are almost as fast as Bitcoin Lightning Network transactions, which are the fastest transactions in the world. Just as well, ChatGPT prompts are so fast that users can get replies in a few seconds for straightforward queries. So while superhuman digital intelligence zips across the world in seconds to make our minds smarter, digital capital zips across the world in seconds to quench wealth asymmetries and provide opportunity to us all. Fast money and Fast intelligence. My kind of speed. *** Miscellaneous: Two whales to inspire smarter smart contracts Given what is happening in the AI and Cryptocurrency/Bitcoin space, future Rootstock smart contracts would do well to peg to a fast-growing Bitcoin whale like MicroStrategy, or a technology whale like Nvidia. Why? Here’s Nvidia’s stock graph in October. And here’s MicroStrategy’s stock graph. MSTR did better than NVDA last month, but they are both going to the moon in the long run. Just like the Bitcoin and AI markets whose CAGR is looking at double digits all the way to 2030 (>20% for both as per Statista.com). For their sake, Rootstock smart contracts will be on these rocket rides. Then they just might become really smart smart contracts. I recently Googled And IBM.com gave the top result. As it explains, smart contracts automate coded actions on a cryptocurrency platform (blockchain). So I’m thinking trader bots and escrow agreements. The result also says they can automate a workflow. So hourly, weekly, or monthly (cryptocurrency) payments to a payroll. OK. I guess there is something smart about smart contracts. Coz people usually handle the aforementioned tasks. Then I got to thinking, can smart contracts be as smart as artificial intelligence algorithms? Yes, I Googled. It’s not a thing. Every site talks of a smart contract combined with AI. I want it to become AI. combined become I know, we all want things. If wishes were horses. Nonetheless, I have a few quirky similarities to share. Starting with the cover image. Notice how the Rootstock logo looks a lot like an Artificial Neural Network? I did. Bitcoin’s supply mechanics and the Boltzmann distribution Rootstock is a Bitcoin side chain which means at its heart, it is Bitcoin’s supply mechanics that govern the halving and the difficulty adjustment and more. The true engine of all its smart contracting. While Rootstock was chugging along its Bitcoin rails, on the other side of town, the Nobel Prize in Physics was being awarded to John Hopfield and Geoffrey Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks.”- https://www.nobelprize.org/ . “for foundational discoveries and inventions that enable machine learning with artificial neural networks.”- https://www.nobelprize.org/ Their work in machine learning led to GPTs (Generative Pre-trained Transformers), a now famous architecture in major AI systems today. What’s interesting is that foundational to all Artificial Neural Networks (ANNs), as mentioned in the Advanced information section of NobelPrize.org, is something called the Boltzmann Distribution. Advanced information Below is its mathematical equation to capture the energy E of an ANN; energy Notice the similarity to the Bitcoin supply equation below (with 3 parameters too! Replace W_ij above with 210,000, S_j with 50, and S_i with 1/2^i) ; What can I say, there is something Bitcoin-like to AI. Both the Boltzmann distribution and the Bitcoin supply equation model an exponential decay/reduction. The former is generalized for the energy of a closed system with n particles, while the latter models the number of bitcoins available to be mined. This too is a closed system with n = number of halvings. The particles can be as many as they want to be (so whether Bitcoin is a planet-sized computational network or a modest network of a dozen PCs, doesn’t matter). But they will all decay in 32 halvings to mine all the bitcoins that can possibly be mined. Approximately 21 million. Now bitcoins are mined in a system that unlike standard computer systems, where linguistic logic coded by a programmer dictates the state of 0s and 1s, this machine changes when it is fed energy... - Sydney Bright, BitcoinMagazine. unlike standard computer systems, where linguistic logic coded by a programmer dictates the state of 0s and 1s, this machine changes when it is fed energy ... - Sydney Bright, BitcoinMagazine . energy BitcoinMagazine How smart are Rootstock smart contracts? We have seen that both Rootstock smart contracts and AI algorithms consume energy. So the question is, how well do they consume the energy? Well, both consume it pretty well to make some good money. They are both smart money algorithms in that regard. To see an instance of this, let’s put Rootstock next to ChatGPT in a certain way. Look at the below statistics. Of course, ChatGPT has more users, more money flowing in, and more employees worldwide. It is smart. We all know this. But pound for computational pound, Rootstock does hold its own. Let’s make some assumptions like a Physicist. For example, we can see that Rootstock.io has 261,406 monthly transactions. Assume they are from 261,406 different people. Meanwhile, ChatGPT has 800 million monthly users, generating billions of prompts (similar to transactions) every month. Let’s take their number of monthly users as equivalent to their monthly transactions. Now let’s divide market cap by these monthly figures (this is a made-up metric, but a good one): Rootstock’s “ Market cap per monthly user ” = $380,000,000 / 261,406 = $1453 Market cap per monthly user ChatGPT’s “ Market cap per monthly user ” = $157,000,000,000 / 800,000,000 = $196 Market cap per monthly user As you can see, ChatGPT is not doing better than Rootstock in this calculation! And if those 2740 RBTC pegged to BTC soon climb to $100k+ per RBTC (courtesy of the Trump Administration), you bet, ChatGPT will learn all about it. And OpenAI just might consider allying with a newly-unicorn Rootstock to provide paid ChatGPT services in RBTC. Bottom line similarity: Both Rootstock smart contracts and ChatGPT are smart money algorithms with good performance metrics. The former caters to a smaller clientele (Bitcoiners) with deep pockets, while the latter caters to a larger clientele (the world’s literate population) with relatively modest pockets. Speedsters In the Rootstock white paper , while the average confirmation time of BTC transactions is 10 minutes, Rootstock’s RBTC, which holds a 1:1 on-chain peg to BTC, takes only 30 seconds in average confirmation time. Rootstock white paper This implies that RBTC transactions are almost as fast as Bitcoin Lightning Network transactions, which are the fastest transactions in the world . fastest transactions in the world Just as well, ChatGPT prompts are so fast that users can get replies in a few seconds for straightforward queries . users can get replies in a few seconds for straightforward queries So while superhuman digital intelligence zips across the world in seconds to make our minds smarter, digital capital zips across the world in seconds to quench wealth asymmetries and provide opportunity to us all. Fast money and Fast intelligence. My kind of speed. *** Miscellaneous: Two whales to inspire smarter smart contracts Given what is happening in the AI and Cryptocurrency/Bitcoin space, future Rootstock smart contracts would do well to peg to a fast-growing Bitcoin whale like MicroStrategy, or a technology whale like Nvidia. Why? Here’s Nvidia’s stock graph in October. And here’s MicroStrategy’s stock graph. MSTR did better than NVDA last month, but they are both going to the moon in the long run. Just like the Bitcoin and AI markets whose CAGR is looking at double digits all the way to 2030 (>20% for both as per Statista.com). For their sake, Rootstock smart contracts will be on these rocket rides. Then they just might become really smart smart contracts. really smart

Stories, News & Mentions about SUPERHUMAN

Superhuman’s Stories on HackerNoon

Superhuman's News & Mentions