Related posts

Venture capitalist and founder of Theory Ventures, previously a partner at Redpoint. Writes a near-daily blog on startups, SaaS metrics, data and AI.

• Our readingkorrents.com

  Safely routing AI tasks to cheaper models requires accumulated observation of many prior task outcomes.

  Their words

  Routing to a cheap model is only safe if you know which tasks it clears, & that knowledge comes from watching ten thousand versions of the same work.

  The Harness Margin Opportunitytomtunguz.com

Dutch bootstrapper who builds and runs their internet startups alone — Nomads.com, Remote OK, Hoodmaps, Photo AI and Interior AI — and has written about it at levels.io since 2013. Posts as @levelsio.

• Our readingkorrents.com

  Europe should ease immigration from countries whose prior immigrants have low crime rates and largely ban inflow from countries with high rates, with yearly data adjustments.

  immigrationEurope

  Their words

  I think we should use this kinds of data to completely turn around Europe's immigration policy: - We should have super easy inflow for immigrants from countries whose previous immigrants have shown low crime rates You're from Japan? Just fly here and you can stay forever - We should then almost completely ban any inflow from countries whose previous immigrants have shown high crime rates with instant deportations

  @levelsio on Xx.com

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

• Our readingkorrents.com

  Deep properties of a model are inherited through the previous generation's data, which is why AI systems from different companies end up correlated with one another.

  Their words

  And so there's some like deep underlying properties of the model that are being sort of transferred between model generations because basically you you train your AI on data from the prior generation and keep going.

  youtube.com

Computer science professor at Georgetown University and author of Deep Work, Digital Minimalism and Slow Productivity; writes about focus, technology and work at calnewport.com and hosts the Deep Questions podcast.

• Our readingkorrents.com

  The personal computer made countless individual tasks easier without delivering the productivity gains everyone expected, and AI may go the same way.

  Their words

  it’s also true that their impact on our output was never as tremendous or as clear-cut as that of prior innovations, such as the steam engine or the power loom, and that integrating them effectively into the workforce took a long time

  Beware of Productivity Paradoxescalnewport.com

Professor of economics at George Mason University and Bartley J. Madden Chair at the Mercatus Center. Co-writes the blog Marginal Revolution and co-authors the textbook Modern Principles of Economics with Tyler Cowen.

• Our readingkorrents.com

  Corner allocations under price ceilings are a distinct source of cross-market misallocation, separate from Harberger quantity loss and within-market misallocation.

  Their words

  These corner allocations create a distinct source of cross-market misallocation, separate from the aggregate quantity loss (the Harberger triangle) and from within-market misallocation emphasized in prior work.

  arxiv.org

Co-founder of Superlogical, started in 2026 to build server-side terminal infrastructure; creator of Ghostty. Co-founded HashiCorp and created Vagrant and Terraform before that.

• Our readingkorrents.com

  Chatbot interfaces are the wrong tool for meaningful coding work, because you are mostly hoping the model recalls the right answer and correcting it requires a human in the loop.

  OpenAIGoogle

  Their words

  Immediately cease trying to perform meaningful work via a chatbot (e.g. ChatGPT, Gemini on the web, etc.). Chatbots have real value and are a daily part of my AI workflow, but their utility in coding is highly limited because you're mostly hoping they come up with the right results based on their prior training, and correcting them involves a human (you) to tell them they're wrong repeatedly.

  My AI Adoption Journeymitchellh.com

Economist; Professor of the Practice of Economic Policy at Harvard, nonresident senior fellow at the Peterson Institute, and Chairman of the US Council of Economic Advisers from 2013 to 2017.

• Lovedrcmnd.app

  After the Spike: Population, Progress, and the Case for People

  Their words

  It had the best opening sentence of any nonfiction book I read this year: "In 2012, 146 million children were born. That was more than in any prior year. It was also more than in any year since. Millions fewer will be born this year. The year 2012 may well turn out to be the year in which the most humans were ever born—ever as in ever for as long as humanity exists." Such a simple and obvious fact but I had not known it before. I had to keep reading.

  fivebooks.com

Co-founder and chief scientist of Safe Superintelligence Inc., and previously co-founder and chief scientist of OpenAI.

• Our readingkorrents.com

  That people are good at maths and coding, which evolution never selected for, shows human learning is not a complicated evolutionary prior but something more fundamental.

  Their words

  What I meant to say is that language math and coding and especially math and coding suggests that whatever it is that makes people good at learning is probably not so much a complicated prior but something more some fundamental thing.

  youtube.com

Head of growth at Chess.com, previously growth and monetization lead at Grammarly and Duolingo and a product manager at Google.

• Our readingkorrents.com

  Doing PR before you have evidence customers want the thing is risky, because the publicity creates a sunk cost that makes the team see a failing product through.

  Their words

  And the third one is like we did a lot of actually PR uh prior to the service going out just to get the word out. Um and you know PR has has its time and place. Uh but I think doing it before you have validation that customers definitely want the thing is quite risky. It can lead to a lot of sunk cost uh once you get it out because you're you're just you know you need to see it through.

  youtube.com

Reinforcement-learning researcher; co-wrote the field's standard textbook, wrote The Bitter Lesson, and shared the 2024 Turing Award.

From one piece Richard Sutton – Father of RL thinks LLMs are a dead end 2 beliefs, in the piece's order there

• Our readingkorrents.com

  Large language models are a bad way to do science, because they have been fed so much that nobody knows what they already knew.

  LLMs

  Their words

  We don't we don't really know what information they had prior. We are we have to guess because they've been fed so much. This is one reason why they're not a good way to do science. Uh it's just so uncontrolled, so unknown.

  youtube.com

• Our readingkorrents.com

  Language models cannot be a prior for learning from experience, because a prior needs a ground truth to be a prior about and they have none.

  Their words

  So there's no ground truth. You can't have prior knowledge if you don't have ground truth because the prior knowledge is supposed to be a hint or an initial belief about what the truth is.

  youtube.com

Founder of Terraform Industries and formerly a physicist at NASA's Jet Propulsion Laboratory; writes about energy, space and manufacturing.

• Our readingkorrents.com

  Every industrial revolution is the same move — routing around a bottleneck — and the bottleneck before ours was the limit of what a human or a horse could digest.

  Their words

  And the bottleneck prior to what we call the industrial revolution was uh metabolism, right? It's just like how much like oats can a human or a horse physically digest and then convert into useful mechanical output for you know their their um their peasant overlord or whatever.

  youtube.com

Energy and decarbonisation analyst; publishes an annual chart deck on the state of the transition, and was chief content officer at BloombergNEF.

• Our readingkorrents.com

  Late-summer solar generation typically takes a larger share of global electricity demand than the prior spring's peak share.

  solarenergy

  Their words

  Shoulder generation (in late summer) of any given year is generally a higher share of global electricity demand than the peak generation (spring) of the year prior.

  Growth Measurementsnathanielbullard.com

Co-founder of Sundial; former vice president of product design at Facebook; author of The Making of a Manager; writes The Looking Glass.

• Our readingkorrents.com

  ChatGPT exploded because of its obvious chat interface, not its technology, which had launched earlier in a different wrapper that few noticed.

  OpenAI

  Their words

  The technology powering ChatGPT had already launched prior to November 30th, 2022. It was just wrapped in a different interface. Very few people paid attention. Why did ChatGPT explode in popularity? Because of its obvious chat interface, which everyone intuitively already knew how to use it.

  lg.substack.com

Co-founder of Foundry Group and of Techstars. Has blogged at Feld Thoughts since 2004 and has written or co-written nine non-fiction books on startups and venture capital.

• Recommendsrcmnd.app

  Ecosystem Arabia: The Making of a New Economy

  Their words

  I have written forewords for Amir in the past on two of his prior books, Ecosystem Arabia: The Making of a New Economy and Venture Adventure: Startup Fundraising Advice from Top Global Investors, both of which I recommend.

  feld.com

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.

American science-fiction writer, author of the collections Stories of Your Life and Others and Exhalation, and an essayist on artificial intelligence for The New Yorker.

• Our readingkorrents.com

  A work does not have to be unlike everything before it to be worth making; what makes it new is that this particular person made it.

  Their words

  What you create doesn’t have to be utterly unlike every prior piece of art in human history to be valuable; the fact that you’re the one who is saying it, the fact that it derives from your unique life experience and arrives at a particular moment in the life of whoever is seeing your work, is what makes it new.

  Why A.I. Isn't Going to Make Artnewyorker.com

Statistician at Carnegie Mellon; writes on complex systems, machine learning and inference, and has kept a books-read notebook since 2000 that is now thousands of entries long.

• Likedaffiliate linkrcmnd.app

  Star Kingdom

  Their words

  Buroker is a reliable author of very fluffy mind candy, either science fiction or fantasy. A lot of her output is re-working material from identifiable prior art, without hewing so closely to the canon as to constitute fan fiction.

  bactra.org

Economist at the University of Houston; wrote Fully Grown, on why a slow-growing economy is a sign of success, and blogs at Growth Economics.

• Our readingkorrents.com

  Before about 1870 every burst of productivity growth was eaten by a burst of population growth, leaving living standards where they were.

  Their words

  Prior to this inflection point, productivity growth had increased before, but every time had induced a response of higher population growth rates, eroding any gains in per-person living standards.

  Review of DeLong's "Slouching Towards Utopia"growthecon.com

Indonesian software engineer, educator and engineering manager at Zero One Group; open-source enthusiast writing at ripandis.com.

• Mixed onrcmnd.app

  13" MacBook Pro 2019

  Their words

  Previous development machine. I was using an Intel-based 13" MacBook Pro prior to this and the difference is night and day. Served well for years before upgrading to the M2 version.

  ripandis.com

Co-founder of Netscape and of the venture firm Andreessen Horowitz. Argues at length for building things, mostly in short form.

• Our readingkorrents.com

  Hostility to a new technology comes from people whose power rests on the old one, because the new technology reorders who holds status.

  Their words

  And so all of the people who have achieved positions of power and authority in the world based on prior generations of technology are inherently threatened by new technology because new technology will upend the power and status structures and then a new set of people will, you know, come to the foreground and and take the power.

  youtube.com

Co-founder and CTO of Oxide Computer Company. Created DTrace at Sun Microsystems and was CTO of Joyent; blogs about systems software, open source and engineering culture.

• Lovedrcmnd.app

  Selma

  Their words

  I had taken my kids to see DuVernay's superlative Selma years prior, and I had heard about 13th (and I had recalled my wife saying she wanted to watch it), but it was – like many things – "in the queue."

  bcantrill.dtrace.org

Co-founder and CEO of Anthropic; previously VP of research at OpenAI.

• Our readingkorrents.com

  Scaling up language models greatly improves task-agnostic few-shot performance, sometimes matching prior fine-tuning approaches.

  Their words

  Here we show that scaling up language models greatly improves task-agnostic, few-shot performance, sometimes even reaching competitiveness with prior state-of-the-art fine-tuning approaches.

  arxiv.org