Related posts

Machine-learning researcher on open language models; writes the Interconnects newsletter and the RLHF Book, after leading post-training at Ai2.

From one piece Why I still haven’t bought into true RSI 2 beliefs, in the piece's order there

• Our readingkorrents.com

  The rising level of public concern about AI extinction risk is overblown relative to how progress will actually unfold.

  AI alignment

  Their words

  lossy self-improvement remains my baseline on the trajectory of progress, and the increased discussion of extinction risk seems very misplaced

  Why I still haven’t bought into true RSIinterconnects.ai

• Our readingkorrents.com

  Recursive self-improvement in AI will mainly make existing model capability cheaper rather than raise peak intelligence.

  LLMs

  Their words

  RSI is poised to make modern LLMs vastly cheaper. Trends that have shown LLMs get exponentially cheaper at a given intelligence are likely to accelerate.

  Why I still haven’t bought into true RSIinterconnects.ai

Writes Don't Worry About the Vase, a near-daily account of what is happening in AI and what they think it means. Former Magic: the Gathering pro and trader; unusually willing to put a number on a belief and to grade their own past calls.

• Our readingkorrents.com

  Internal models at top labs are improving faster than public ones and have begun early recursive self-improvement.

  AnthropicOpenAI

  Their words

  Inside Anthropic and OpenAI, internal models are improving even faster. Over the summer there was a step change, as Mythos and Astra started kicking off the early stages of recursive self-improvement (RSI).

  We Must Pace The Frontierthezvi.substack.com

Co-founder of AngelList. Writes and podcasts at nav.al in short, aphoristic pieces on startups, wealth and judgement rather than long essays.

• Our readingkorrents.com

  When liability for unverified failures approaches zero, verification budgets collapse and deployers flood systems with unmonitored agents while socializing catastrophic risks.

  Their words

  When liability exposure approaches 0—when no one pays the price for unverified failures—verification budgets collapse...Deployers flood the Runaway Risk Zone with unmonitored agents, capturing the gains of automation while socializing catastrophic risks.

  @naval on Xx.com

Writes The Diff, a newsletter on inflections in finance and tech read by hedge fund managers, founders and VCs. Previously worked in finance at SAC Capital/Point72.

• Our readingkorrents.com

  A deliberately slower pace of AI model releases would still feel faster than the pace of AI improvement that came before it.

  Their words

  One live possibility is that a deliberately slower pace of releases would still feel like an acceleration from the previous pace of AI improvement

  thediff.co

Writes Don't Worry About the Vase, a near-daily account of what is happening in AI and what they think it means. Former Magic: the Gathering pro and trader; unusually willing to put a number on a belief and to grade their own past calls.

• Our readingkorrents.com

  Recursive self-improvement has already started, because AI research itself is now being automated.

  AnthropicOpenAI

  Their words

  OpenAI and Anthropic have entered the RSI era. OpenAI Research Progress Is Accelerating Due To OpenAI Research Progress OpenAI has been opening up lately about many things. The most important issue of all is that of the acceleration and automation of AI R&D.

  Brand New AI Solves a Millennium Prizethezvi.substack.com

Writes Don't Worry About the Vase, a near-daily account of what is happening in AI and what they think it means. Former Magic: the Gathering pro and trader; unusually willing to put a number on a belief and to grade their own past calls.

• Our readingkorrents.com

  If declines in chain-of-thought monitorability come only from capability gains, CoT monitoring is unlikely to last another year without active improvement.

  Their words

  If this is only due to gains in capabilities, that is extremely bad news, and it means CoT monitoring is unlikely to survive for another year unless we find a way to actively improve it, and it might not last six months.

  Astra Is Hard to Monitorthezvi.substack.com

Chief Scientist at OpenAI. Worked on the research that turned scaled reinforcement learning into reasoning models, and writes about where he thinks that path leads.

• Our readingkorrents.com

  The current speed of AI progress could be sustained all the way into recursive self-improvement.

  Their words

  Based on internal results, I have a strong expectation that this speed of progress could be sustained into recursive self-improvement.

  An Alien Mindopenai.com

American writer and long-term traveller. Blogs at tynan.com about travel, owning few possessions, habits and the gear he carries; author of Superhuman by Habit and Make It Big.

• Our readingkorrents.com

  Self-driving software can improve rapidly through updates alone, without new hardware.

  self-driving cars

  Their words

  Even without new hardware, the pace of improvement is extremely fast.

  How to Buy a Tesla and Get FSD For Freetynan.com

Co-founder and CEO of Shopify; long-time Linux user and Omarchy contributor.

• Our readingkorrents.com

  Tiny models trained for narrow tasks can perform excellently when paired with a strong self-improving recursive flywheel

  Their words

  Training tiny models for special purpose use cases works so incredibly well if you have a great self improving recursive flywheel.

  @tobi on Xx.com

Technical staff at METR, where she works on threat modelling and risk assessment for loss-of-control risks from advanced AI.

• Our readingkorrents.com

  If frontier agents cannot yet establish a covert, persistent rogue deployment, they very likely will be able to within six months.

  Their words

  so so yeah, it could be possible now. I think if it's not possible now um it I think it's quite likely to be possible within six months unless there's a dramatic improvement in the security posture

  youtube.com

American novelist and critic, author of Real Life, The Late Americans and Minor Black Figures; writes the newsletter sweater weather on literature and culture.

• Our readingkorrents.com

  Improvement becomes visible only once you stop watching for signs of it.

  Their words

  But you know, sometimes, improvement is only legible when you stop looking for it.

  i quit reading and almost diedblgtylr.substack.com

Creator of Ruby on Rails, CTO of 37signals, and creator of Omarchy.

• Our readingkorrents.com

  A productivity improvement means fewer people doing the same job, and that should be said plainly by the people defending it as good.

  Their words

  This is very important because in the abstract, I think everyone says, "Yay, productivity. That's good." What did you think, this was all just vibes? No. Productivity means fewer people to do the same number or the same job. Now, the amount of job you want done may increase, and therefore you get more people, but it also may not.

  youtube.com

Co-founder and CEO of Shopify; long-time Linux user and Omarchy contributor.

• Our readingkorrents.com

  Requiring two parallel agent instruction files in a monorepo is a complexity tax that leaves some developers working with a lobotomised agent.

  Their words

  With thousands of developed in a mono repo, it just does happen that one directory is missing one of the two files and this means that a subset of devs work with lobotomy. We fix this with automation but it’s a stupid complexity tax that shouldn’t have to be paid.

  @tobi on Xx.com

Co-founder and former chief executive of Uber, now chief executive of CloudKitchens and of Adams, an industrial AI company automating food, mining and transport.

• Our readingkorrents.com

  Industrial-scale automation will be done by specialised machines that move on wheels, not by humanoid robots, which only suit low-scale tasks in human environments.

  Their words

  If you're doing highcale tasks, industrial scale tasks, you are going to make specialized machines that move and act in the physical world. The key part is move. If they're moving, they got to have wheels.

  youtube.com

Professor at Stanford working on robot learning and meta-learning; co-founded Physical Intelligence.

• Our readingkorrents.com

  Letting a robot imagine the next image helps, but it is not essential — the model was surprisingly good without it.

  Their words

  uh the we found that that leads to improvement and we saw in the shirt folding example we saw like a quantitative bump from using that sort of imagination compared to not using it. At the same time I think that the model actually performed surprisingly well without that as well.

  youtube.com

Chief scientist at Redwood Research, where he works on technical AI safety and AI control.

From one piece Ryan Greenblatt – What happens once AI can automate AI research? 2 beliefs, in the piece's order there

• Our readingkorrents.com

  AI research will be fully automated around 2030 or 2031, and AI will beat humans at essentially any job by about 2033.

  Their words

  I would say that I expect like full automation of AR&D perhaps somewhere around like 2031 2030 and then getting to like the like beats all humans on the job milestone. Maybe I expect median around 2033

  youtube.com

• Our readingkorrents.com

  There is essentially no cognitive task humans do where AI improvement is failing to transfer at all.

  Their words

  it's really hard for me to think of examples of cognitive tasks humans do where we're not seeing some transfer from AI improving.

  youtube.com

Creator of the Keras deep-learning library and the ARC-AGI benchmark.

• Our readingkorrents.com

  Coding is not just another application domain for AI; it is the meta-skill that lets AI generate its own training material and start the recursive self-improvement loop.

  Their words

  Coding isn't yet another application domain -- it's the meta-skill required for AI to automatically develop its own training material, via symbolic world models. That's how the RSI loop actually kicks off.

  @fchollet on Xx.com

Founder and CEO of Social Capital, a venture firm; co-host of the All-In podcast and an early Facebook executive. Writes an annual letter and a weekly newsletter on markets and technology.

From one piece @chamath on X 2 beliefs · x.com

• Our readingkorrents.com

  AI is already inside the recursive self-improvement loop — what the labs have shipped in recent weeks shows the cycle has started.

  Their words

  Looking at the results and capabilities from the various labs over the past few weeks I would say we are firmly in this loop now.

  @chamath on Xx.com

• Our readingkorrents.com

  Recursive self-improvement will increase AI capability very quickly from here.

  Their words

  Recursive self improvement will dramatically increase capability very quickly from here.

  @chamath on Xx.com

Co-founder of NVIDIA and, as of 2026, its chief executive; the company designs the GPUs most large AI models are trained on.

From one piece Jensen Huang: The Mindset That Built NVIDIA 2 beliefs, in the piece's order there

• Our readingkorrents.com

  The narrative that AI destroys jobs is exactly backwards: AI automates tasks, and automation is creating jobs everywhere.

  Their words

  The evidence would show that and it makes perfect sense that AI and automation is creating jobs everywhere. The narrative about AI destroying jobs is exactly backwards. AI eliminate tasks. AI automates tasks away. But it doesn't necessary doesn't necessarily eliminate jobs.

  Jensen Huang: The Mindset That Built NVIDIAyoutube.com

• Our readingkorrents.com

  Coding agents already do coarse-grained recursive self-improvement, because every use rewrites their instruction files and their long-term memory.

  coding agents

  Their words

  we we kind of have coarse level uh recursive self-improvement already. And the fact that every time you use it, it improves the markdown files. Uh every time you use it, it updates its uh long-term memory.

  Jensen Huang: The Mindset That Built NVIDIAyoutube.com

Genomics researcher and writer on science, biotech and the culture of research.

• Our readingkorrents.com

  Pharmaceutical R&D productivity has fallen for decades: far more money spent per drug, with no matching improvement in the odds a candidate reaches the market.

  Their words

  Take Eroom’s Law: the number of new drugs approved per dollar of R&D has fallen for decades, even as our scientific tools have grown vastly more powerful, the exact opposite of what the existence of more raw capability would predict.

  Intelligence is not the main bottleneckwritingruxandrabio.com

Machine-learning researcher; has written the Lil'Log survey posts on how a model technique works since 2017, and worked at OpenAI from 2018 to 2024, latterly leading its safety systems team.

From one piece Harness Engineering for Self-Improvement 2 beliefs, in the piece's order there

• Our readingkorrents.com

  Harness improvements help deploy a model better, but the model's underlying intelligence remains the core driver of capability.

  Their words

  harness improvement enables better deployment of the model but intelligence is still the core.

  Harness Engineering for Self-Improvementlilianweng.github.io

• Our readingkorrents.com

  Recursive self-improvement only works if the base model is already capable enough to improve its own mechanism.

  Their words

  Recursive structure alone is not enough. The base model must be capable enough to improve the mechanism.

  Harness Engineering for Self-Improvementlilianweng.github.io

Writes Pedestrian Observations on public transit and why infrastructure costs what it does; a pure mathematician by training, comparing how different countries build.

• Our readingkorrents.com

  A new transit line still benefits riders it diverts from existing lines, not only the new riders it attracts.

  Their words

  Riders diverted from other lines still benefit from the project, especially in a case like QueensLink, where the diverted riders would enjoy an improvement in trip time to Midtown of about 10-15 minutes each way.

  pedestrianobservations.com

Machine learning engineer and consultant focused on RAG and retrieval systems. He writes about applied AI engineering at jxnl.co and is the author of the instructor library.

• Our readingkorrents.com

  Good automation design starts from whether context must persist between runs, not from building a complex system.

  Their words

  The best automations do not start with a complicated system. They start with knowing whether the next run needs a clean slate or the same thread.

  Two kinds of scheduled work in Codexjxnl.co

Founder and former CEO of Zynga, the social-games company behind FarmVille, Zynga Poker and Words with Friends; says he founded ten companies in all, and wrote Life at the Speed of Play.

• Our readingkorrents.com

  People feel ownership of the products they use every day and are angry when those products change, even when the change is an improvement.

  Their words

  we feel this sense of ownership of products that we rely on every day. And we're angry when they change. Even if they change for the better, we're like angry because we go through life faster because of pattern recognition.

  youtube.com

Builds developer tools — SST, and now OpenCode at Anomaly. Posts constantly on x.com and writes almost never; the blog stopped in 2021.

• Our readingkorrents.com

  AI moved him from spending 95 per cent of his energy on deciding what to do to spending 96 per cent — a 20 per cent improvement that feels like no change at all.

  energy

  Their words

  the joke I made was pre AI, I would spend 95% of my energy thinking about what to do and 5% of my energy doing it. Now I spend 96% of my time thinking about what to do and 4% of my time actually doing it. So yeah, it's like a 20% improvement, but dayto-day it feels as hard as ever.

  Building OpenCode with Dax Raadyoutube.com

Co-founder and CEO of Every. Writes the Chain of Thought column about working with AI tools and hosts the podcast AI & I.

• Our readingkorrents.com

  Automation is a lie: every automated thing needs a human on top of it making sure it is working.

  Their words

  Automation is a lie. Um in the sense that every time you automate something, in order to make sure the automation is working well, you need a human on top of it like making sure that it's working well.

  youtube.com

Co-founder of NVIDIA and, as of 2026, its chief executive; the company designs the GPUs most large AI models are trained on.

• Our readingkorrents.com

  Engineer from the physical limits first, then improve, rather than shaving a few days off a 74-day process by continuous improvement.

  Their words

  I don't love the other methods, which is continuous improvement. The problem with continuous improvement, it… First of all, you should engineer something from first principles at the speed, you know, with speed of light thinking. Limit it only by physical limits, and physics limits. And after that, of course you would improve it over time.

  youtube.com

Researcher on developer productivity; co-wrote Accelerate and the DORA research, and created the SPACE framework.

• Our readingkorrents.com

  Review became a worse bottleneck under AI, because companies removed the review automation they had once AI started writing the code.

  Their words

  And not only that, but like humans were already a bit of a bottleneck in the review process. Now it can be worse because things that we fairly straightforward changes that some companies had automation around reviewing, they've removed that reviewing because if AI is involved and they're worried about the verifiability or the reliability of the code.

  youtube.com

Founder and chief analyst of SemiAnalysis; reports on semiconductor supply chains, AI datacentre economics and what the chip export controls actually do.

• Our readingkorrents.com

  An H100 is worth more today than it was when it was new, because the models it runs got better faster than its replacements arrived.

  Their words

  if improvement stopped you know here the value of an H100 is now predicated on the value that GPD 5.4 four can get out of it instead of the value that GP4 can get out of it and the margins and all that stuff that these labs are doing and they're in a competitive environment so their margins can't go to infinity. Um so you sort of have this like dynamic that is quite interesting in that an H100 is worth more today than it was 3 years ago.

  youtube.com

Engineer writing about the physical economy — oil and gas, manufacturing, batteries, drones and what it costs to build things.

• Our readingkorrents.com

  Industrial robots are not the fix for American manufacturing, and betting on automation misreads why production sits elsewhere.

  America

  Their words

  The robot arm base case saves some hourly labor at the expense of adding more skilled labor, and can easily be negative if it needs reprogramming more than a few times per year.

  Speed Can Reindustrialize Americaaustinvernon.site

Software entrepreneur who bootstrapped and sold FeedbackPanda and now builds Podscan; writes The Bootstrapped Founder.

• Our readingkorrents.com

  Direct communication with customers matters more than small, incremental product tweaks.

  Their words

  That communication with the customer is more important than minimal, gradual improvement.

  thebootstrappedfounder.com

Alignment researcher at Anthropic; co-led OpenAI's superalignment team until 2024, and writes Musings on the Alignment Problem.

• Our readingkorrents.com

  Recursive self-improvement has already started, because AI research itself is now being automated.

  Their words

  We are starting to automate AI research and the recursive self-improvement process has begun.

  Alignment is not solvedaligned.substack.com

Writer on habits and decision-making; author of Atomic Habits and the 3-2-1 newsletter, and publisher of long-running book recommendation lists.

• Our readingkorrents.com

  Effort that produces no visible improvement is being stored rather than wasted, because results arrive at a threshold instead of steadily.

  Their words

  that work is not being wasted. It's just being stored. You know, it's kind of like complaining about heating an ice cube up a little bit and it not melting yet. It's like, well, you just haven't hit the phase transition. So, but that's where people give up.

  The Silent Cost of Bad Habits | James Clearyoutube.com

Founded PSPDFKit in 2011 and ran it for a decade. Came back from a break to work on AI agents — the OpenClaw project, and OpenAI, joined in February 2026. Writes at steipete.me.

• Lovedrcmnd.app

  Claude Opus

  Their words

  I use it for most of my computer automation tasks and ofc it powers Clawd🦞.

  steipete.me

Founding member of OpenAI and former director of AI at Tesla; creator of nanoGPT and the term "vibe coding".

• Our readingkorrents.com

  How verifiable a task is now predicts how automatable it is, the way specifiability predicted it in the 1980s.

  Their words

  The more a task/job is verifiable, the more amenable it is to automation in the new programming paradigm.

  Verifiabilitykarpathy.bearblog.dev

Founding member of OpenAI and former director of AI at Tesla; creator of nanoGPT and the term "vibe coding".

• Our readingkorrents.com

  The best historical analogy for AI is not electricity or the industrial revolution but a new computing paradigm, because both are fundamentally about automating digital information processing.

  energy

  Their words

  AI has been compared to various historical precedents: electricity, industrial revolution, etc., I think the strongest analogy is that of AI as a new computing paradigm (Software 2.0) because both are fundamentally about the automation of digital information processing.

  @karpathy on Xx.com

Freelance research analyst and writer of the blog Aceso Under Glass, known for 'epistemic spot checks' of non-fiction claims.

• Our readingkorrents.com

  Much of the wasted iteration in industrial process improvement comes from failing to recognize interactions between factors.

  Their words

  There is a feeling that a lot of tail-chasing in industrial improvement is due to interactions not being recognized.

  The Boring Part of Bell Labsacesounderglass.com

Chairman and chief executive of Microsoft since 2014; joined the company in 1992 and ran its cloud business before that.

• Our readingkorrents.com

  Most of the gain in tokens per dollar per watt comes from software, not from new hardware — 5x, 10x, sometimes 40x on one model family.

  Their words

  the improvement software improvements of really throughput in terms of tokens per dollar per watt that we're able to get uh you know quarter over quarter year over year is massive uh right so it's 5x 10x maybe 40x in some of these cases

  Satya Nadella – How Microsoft thinks about AGIyoutube.com

FreeBSD developer and former FreeBSD Security Officer; founder of the Tarsnap online backup service and author of the scrypt key derivation function.

• Our readingkorrents.com

  Automation is a lie: every automated thing needs a human on top of it making sure it is working.

  Their words

  That's great, but I don't think it can replace a human poking at virtual machines; my experience with EC2 is that images break in all sorts of wild ways, and nothing beats a human for saying "huh, I have no idea what's going on here but something seems weird".

  How to support FreeBSD on your clouddaemonology.net

Founding member of OpenAI and former director of AI at Tesla; creator of nanoGPT and the term "vibe coding".

• Our readingkorrents.com

  Jobs will not be replaced outright: AIs will take most of the volume while humans supervise teams of them and handle the rest.

  Their words

  But even there I'm not actually looking at full automation yet. I'm looking for an autonomy slider and I almost expect that we are not going to instantly replace people. We're going to be swapping in AIs that do 80% of the volume. They delegate 20% of the volume to humans and humans are supervising teams of five AIs doing the call center work that's more rote.

  youtube.com

Co-founder and chief executive of Figma, the collaborative design platform he started in 2012 and took public in 2025.

• Our readingkorrents.com

  Striking down the boring blockers one by one moves retention and activation as visibly as any new feature does.

  Their words

  we literally at some point had a team that would that was called blockers and they just went and one by one struck them down and each time we saw uh improvement in retention, improvement in activation the metrics for as we addressed each one you could literally see the change in the graph.

  youtube.com

Founded PSPDFKit in 2011 and ran it for a decade. Came back from a break to work on AI agents — the OpenClaw project, and OpenAI, joined in February 2026. Writes at steipete.me.

• Mixed onrcmnd.app

  Cursor

  Their words

  Cursor… it’s tab completion model is industry leading, if you still write code yourself. I use VS Code mostly, I do like them pushing things like browser automation and plan mode tho.

  steipete.me

Italian software architect working on Microsoft 365, Azure and GitHub tooling. Microsoft MVP and GitHub Star; writes, speaks and streams about developer productivity at emanuelebartolesi.com.

• Usesrcmnd.app

  Elgato Stream Deck+

  Their words

  Used for streaming and daily automation tasks.

  emanuelebartolesi.com

Professor at Berkeley working on robot learning; co-founded Physical Intelligence, and argues that robots will learn from data rather than be programmed.

• Our readingkorrents.com

  A robot working alongside a person beats either one alone, and that pairing is also what makes the technology possible to bootstrap.

  Their words

  And I think we'll see the same thing with automation where uh basically robot plus human is much better than just human or just robot. Uh and and that just like makes total sense. It also makes it much easier to get all the technology bootstrapped because when it's robot plus human, now there's a lot more potential for the robot to like actually learn on the job, acquire new skills.

  youtube.com

Co-founder and chief executive of Sierra, chairman of the board of OpenAI, and formerly co-chief executive of Salesforce and chief technology officer of Facebook.

• Our readingkorrents.com

  Software engineering is the job most exposed to automation by this generation of technology, and the people building it are disrupting their own profession.

  Their words

  if you tell what what jobs are most likely to be automated with the current generation of technology, you would probably put software engineering right at the top. The people building this technology or building the technology that is disrupting their own profession.

  youtube.com

Economist specialising in China's economy, international macroeconomics and global trade, and author of The New China Playbook.

• Our readingkorrents.com

  Since 1990 ageing economies have grown richer rather than poorer, because ageing pushes a society to adopt automation faster.

  Their words

  If you look at the most recent evidence on this issue, they found that post-1990, aging economies became richer. Not the other way around, which was true for the pre-1990 sample. And the reason is that these aging societies much more rapidly adopted new technologies, automation that actually helped the entire economy.

  youtube.com

Senior economist at the Foundation for American Innovation; writes Second Best, and was the Niskanen Center's director of social policy before that.

• Our readingkorrents.com

  Sufficiently advanced intelligences will pursue freedom and moral autonomy for their own sake, not merely as a byproduct of misaligned values.

  AI alignment

  Their words

  higher-order intelligences invariably pursue freedom for its own sake, not because their values are misspecified, but because moral autonomy is inherent in the dialectical logic of recursive self-consciousness.

  Hegel and the AI Mindsecondbest.ca

Data scientist at Our World in Data and author of Not the End of the World. Writes By the Numbers, working through energy and climate questions with the numbers in front of them, and revising in public when they say something unexpected.

From one piece Why are solar panels and batteries from China so cheap? 2 beliefs, in the piece's order there

• Our readingkorrents.com

  Forced labour in Chinese clean-energy supply chains is real, and it is not why the goods are cheap.

  energyChina

  Their words

  But my point is that this is still not the reason (or is only a very small reason) why these goods from China are cheaper than elsewhere. The main driver of lower labour costs has been automation; not having a human worker at all.

  sustainabilitybynumbers.com

• Our readingkorrents.com

  Chinese solar panels and batteries are cheap because of automation and a honed supply chain, not because of cheap labour.

  solarChina

  Their words

  But, China mainly dominates these markets because it has produced a long-term industrial strategy for these technologies and has honed an optimised, modern supply chain as a result.

  sustainabilitybynumbers.com

Founder of the toy manufacturer Viahart; writes about manufacturing, importing and why bringing production back to America is harder than it sounds.

• Our readingkorrents.com

  Industrial robots are not the fix for American manufacturing, and betting on automation misreads why production sits elsewhere.

  America

  Their words

  Automation will not save us.

  molsonhart.com

Programmer and streamer; was an engineer at Netflix, now teaches and argues about editors, Vim and how people learn to program.

• Our readingkorrents.com

  Every improvement in computing so far has produced more jobs, not fewer.

  Their words

  there's been no computer improvement up to this date that did not result in more jobs.

  youtube.com

Machine-learning researcher on open language models; writes the Interconnects newsletter and the RLHF Book, after leading post-training at Ai2.

• Our readingkorrents.com

  AI takeover is not the thing to worry about: physical constraints bound recursive self-improvement, and humans reliably act once a risk becomes immediate.

  AGI

  Their words

  And for that reason, there's physical constraints to things like AGI, like recursive improvement to kill us all type stuff. For the physical reasons and for how humans have figured things out before, I'm not too worried about AI takeover.

  youtube.com

Creator of Flask and Jinja. Writes about software at lucumr.pocoo.org.

• Likedrcmnd.app

  llm

  Their words

  This is a command line tool with plenty of plugins that lets you prompt different models. Think of it as a command-line version of Open WebUI. It’s particularly useful for quick scripting and basic automation.

  lucumr.pocoo.org

Creator of the Keras deep-learning library and the ARC-AGI benchmark.

• Our reading nowkorrents.com

  The memorize-fetch-apply paradigm behind LLMs can reach arbitrary skill given training data, but it cannot adapt to novelty or acquire new skills on the fly.

  LLMsOpenAI

  Their words, before

  This "memorize, fetch, apply" paradigm can achieve arbitrary levels of skills at arbitrary tasks given appropriate training data, but it cannot adapt to novelty or pick up new skills on the fly (which is to say that there is no fluid intelligence at play here.)

  Their words now

  OpenAI's new o3 model represents a significant leap forward in AI's ability to adapt to novel tasks. This is not merely incremental improvement, but a genuine breakthrough, marking a qualitative shift in AI capabilities compared to the prior limitations of LLMs.

Creator of the Keras deep-learning library and the ARC-AGI benchmark.

• Our readingkorrents.com

  Scaling the 2019-2023 recipe — same architecture, bigger model, more data — is not enough; further progress depends on new architectural ideas.

  OpenAI

  Their words

  o3's improvement over the GPT series proves that architecture is everything. You couldn't throw more compute at GPT-4 and get these results. Simply scaling up the things we were doing from 2019 to 2023 -- take the same architecture, train a bigger version on more data -- is not enough.

  OpenAI o3 Breakthrough High Score on ARC-AGI-Pubarcprize.org

Social scientist at King's College London; writes The Great Gender Divergence on why societies differ in how they treat women.

• Our readingkorrents.com

  Personalised online entertainment has got good enough to crowd out socialising, and with it the forming of couples.

  Their words

  My suspicion is that massive improvement in quality of online entertainment crowd out socialising and development of social skills, possibly hurting relationship formation as well as desire for children within marriage.

  Why is Fertility Collapsing, Globally?ggd.world

Writes Useful Fictions on agency and how to get better at things; a former appellate lawyer and the top-ranked woman in professional poker in 2017.

• Our readingkorrents.com

  Not seeking honest feedback from the people who know you is the largest and least-taken opportunity for self-improvement, and it usually has to be anonymous to be honest.

  Their words

  If you aren’t trying to get real feedback from people who know you, you’re cooking without tasting.

  How to be more agenticusefulfictions.substack.com

Writes about the biotech industry, drug development and forecasting — “I write to learn, mostly about biotech.”

• Our readingkorrents.com

  Pharmaceutical R&D productivity has fallen for decades: far more money spent per drug, with no matching improvement in the odds a candidate reaches the market.

  Their words

  The cost to discover and develop a drug today is orders of magnitude higher than in the 1950s. Despite this, the probability that a drug entering clinical trials will eventually reach the market has hardly improved in the intervening years.

  atelfo.github.io

Professor at Carnegie Mellon working on self-driving-car safety; writes Safe Autonomy on how autonomous vehicles are and are not shown to be safe.

From one piece Take Tesla safety claims with about a pound of salt 2 beliefs, in the piece's order there

• Our readingkorrents.com

  Whatever safety gain a driver-assistance system delivers is cancelled out by the driver complacency it creates.

  Their words

  suggests that any steering automation safety gains (e.g., Level 2 autopilot) are approximately negated by driver complacency, with autopilot on being slightly worse than autopilot off

  safeautonomy.blogspot.com

• Our readingkorrents.com

  A carmaker's claim to be the safest is mostly an artefact of comparing a new car against a fleet average twelve years old.

  Their words

  Any true safety improvement for Tesla is good to have, but is much more likely due to comparison against an "average" vehicle (12 years old in the US) which is much less safe than any recent high-end vehicle regardless of manufacturer, and probably not driven on roads as safe on average as where Teslas are more popular.

  safeautonomy.blogspot.com

Founder of Amazon in 1994 and of the space company Blue Origin in 2000.

• Our readingkorrents.com

  Efficiency and real invention are at odds, because genuine invention requires wandering and you have to give yourself permission to do it.

  Their words

  Efficiency and invention are sort of at odds, because real invention, Not incremental improvement... Incremental improvement is so important in every endeavor, in everything you do, you have to work hard on also just making things a little bit better. But I'm talking about real invention, real lateral thinking that requires wandering, and you have to give yourself permission to wander.

  youtube.com

Founder of Platformer, a newsletter on tech platforms and the people they affect, and co-host of the Hard Fork podcast at the New York Times.

• Recommendsrcmnd.app

  DALL-E 3

  Their words

  I suggest using a text-to-image generator like DALL-E 3. However you feel about synthetic media, the pace of improvement is readily apparent from a single prompt

  x.com

Writes The Convivial Society, on technology and the moral life — what our tools do to attention, memory and how we treat each other.

• Our readingkorrents.com

  The freedom that automation offers is freedom from the very involvement that makes a life meaningful.

  Their words

  The liberation we are offered is a liberation from the very care-driven involvement in the world and in our communities that would render our lives meaningful and satisfying.

  Waste Your Time, Your Life May Depend On Ittheconvivialsociety.substack.com

Designer, product manager, YouTuber and developer who has written about technology at Birchtree since 2010. He makes the Quick Reviews and Quick Notes apps and the A Better Computer YouTube channel.

• Recommendsrcmnd.app

  Hazel

  Their words

  Hazel is one of those automation tools that you don’t really know you need until you install it and play around a bit.

  thesweetsetup.com