Pranav
82 followers · 36 following · 2274 views
on the atlas — 98
- How La Marzocco Perfected the Espresso Machine | Totei - The Art Is in the Making2 savers
- Cyborgism - LessWrong8 savers
- How to understand modern India - Marginal REVOLUTION2 savers
- The Unreasonable Effectiveness of Recurrent Neural Networks15 savers
- An Apple-Picking Model of AI R&D | Tom Cunningham – Tom Cunningham7 savers
- 2025 letter | Dan Wang62 savers
- Do Thing, Do One Thing49 savers
- How to Stage a Coup - by Santi Ruiz - Statecraft8 savers
- Gaddafi hires 200 young Italian women – to convert them to Islam | Muammar Gaddafi | The Guardian1 savers
- The First Fully General Computer Action Model | blog35 savers
- Tristan's Site - Tristan Hume13 savers
- Why transformative artificial intelligence is really, really hard to achieve13 savers
- Balance of power4 savers
- pistar06.pdf2 savers
- FP²: Fully in-Place Functional Programming2 savers
- On Grand Ambitious Theories6 savers
- [2509.14786] Pre-training under infinite compute1 savers
- The New World - Colossus24 savers
- You wouldn’t be here for someone else - by BDM1 savers
- Hyman G. Rickover3 savers
- Shareholder Letters2 savers
- TensorStream - Perceptron4 savers
- Tensor Considered Harmful2 savers
- Book Bounties6 savers
- Defeating Nondeterminism in LLM Inference - Thinking Machines Lab40 savers
- diffuse.one2 savers
- You Have to Feel It – Mitchell Hashimoto5 savers
- Peter Watson (intellectual historian) - Wikipedia1 savers
- Econ syllabus - Google Docs3 savers
- TPU Deep Dive10 savers
- Infomax - Wikipedia1 savers
- Diffusion is not necessarily Spectral Autoregression | Fabian Falck2 savers
- fate - Nina’s Substack2 savers
- Constitutional AI: Harmlessness from AI Feedback5 savers
- Nintil - On Bloom's two sigma problem: A systematic review of the effectiveness of mastery learning, tutoring, and direct instruction7 savers
- yokochos, a perfect third place - by Chloe - milky cities2 savers
- Karl Popper (Stanford Encyclopedia of Philosophy)5 savers
- Why pessimism sounds smart4 savers
- How to Do Great Work80 savers
- What Should You Do with Your Life? Directions and Advice - Alexey Guzey73 savers
- Why books don't work64 savers
- Here's to the fools who dream54 savers
- Milan Cvitkovic52 savers
- How I've run major projects | benkuhn.net52 savers
- Fast · Patrick Collison48 savers
- 95%-ile isn't that good44 savers
- Making Normal Conversations Better - by Sasha Chapin41 savers
- To Enjoy UWaterloo40 savers
- Andrej's advice for success39 savers
- Godly — The Best Web Design Inspiration36 savers
- 🌻 audience of one - half-baked futures | jasmine sun36 savers
- Searching for outliers | benkuhn.net36 savers
- What I Miss About Working at Stripe - Every33 savers
- On Self-Respect: Joan Didion’s 1961 Essay from the Pages of Vogue | Vogue31 savers
- Making Deep Learning Go Faster29 savers
- Moxie Marlinspike >> Blog >> Career Advice29 savers
- The Rationality Community Sucks28 savers
- My techno-optimism27 savers
- How to Optimize a CUDA Matmul Kernel for cuBLAS-like Performance: a Worklog24 savers
- How do you find your people, when searching is itself an antipattern? | Kevin Liu20 savers
- On caring19 savers
- nothing stops18 savers
- The housing theory of everything - Works in Progress17 savers
- Transformer Inference Arithmetic | kipply's blog17 savers
- Don't Major in Computer Science - by Varun Shenoy15 savers
- Just know stuff. (Or, how to achieve success in a machine learning PhD.) · Patrick Kidger14 savers
- The maze is in the mouse. What ails Google. And how it can turn… | by Praveen Seshadri | Feb, 2023 | Medium14 savers
- On Oppenheimer and Progress14 savers
- Think real hard | benkuhn.net13 savers
- There’s more to mathematics than rigour and proofs | What's new13 savers
- Diffusion is spectral autoregression – Sander Dieleman13 savers
- washingtonpost.com: Unabomber Special Report12 savers
- Guessing the Teacher's Password12 savers
- Home | nand2tetris11 savers
- DALL·E 211 savers
- Galactic Empire | Purdue Hackers11 savers
- Growing Neural Cellular Automata10 savers
- The mystery of the miracle year - by Dwarkesh Patel10 savers
- Hire people who give a shit. - by Alexandr Wang10 savers
- A Short History of My Last Six Years — Wait But Why9 savers
- Neuroscience is pre-paradigmatic. Consciousness is why9 savers
- Optimizing AI Inference at Character.AI9 savers
- how I think when I think about programming - alice maz9 savers
- ⚡️ Take Back the Future! - Reboot8 savers
- I am asking you to want something - by Gonzalo Nuñez7 savers
- The Zero-Day Flaw in AI Companies7 savers
- blog/2024.09.impact.md at main · okhat/blog6 savers
- How do trusted setups work?6 savers
- PyTorch internals : Inside 245-5D6 savers
- Neural Networks, Types, and Functional Programming -- colah's blog6 savers
- therapy as a way of aligning with your subconscious5 savers
- i was half myself without you, now i feel so complete5 savers
- Beware the Metagame5 savers
- Pain4 savers
- What is Category Theory Anyway?3 savers
- Replit - Productizing Large Language Models3 savers
- crushes are often just misplaced ambition - by Isabel2 savers
- One Div Zero: A Brief, Incomplete, and Mostly Wrong History of Programming Languages2 savers
highlights — 67
Labs cannot build competitive wrappers because no single lab has homogeneously had the best foundation model.
The Zero-Day Flaw in AI CompaniesWhimsy is in much shorter supply than ambition. Find people to have wholesome fun with and your life will improve greatly.
To Enjoy UWaterlooThere's a kind of excited curiosity that's both the engine and the rudder of great work. It will not only drive you, but if you let it have its way, will also show you what to work on.
How to Do Great WorkIt is our current historical juncture, of the insecurity about a faltering Pax Americana, that Oppenheimer has captured—the nostalgia for a moment in which America’s technological hegemony was consolidated, however briefly.
On Oppenheimer and ProgressNumber one: we can find a way to make people liable for causing catastrophe
How to Prevent the Next Pandemic—Asteriskat least 30,000 people can assemble an influenza virus from scratch
How to Prevent the Next Pandemic—Asteriskdidn’t want to live in the ivory tower: products are meant to be used; writing is meant to be read.
🌻 audience of one - half-baked futures | jasmine sunutopian demand is not informed by political feasibility alone but by its end vision
⚡️ Take Back the Future! - RebootBy the time diversity, ethics, and access enter the conversation, large corporations will have already locked in their technological dominance and naturalized their harms
⚡️ Take Back the Future! - RebootI still strongly believe that one of the best signals for someone who deeply and intrinsically cares about technology is one who can still play and tinker for the hell of it.
Rhizome Research LogThat’s the only way it starts, community, not capital.
Where Cal founders come from | thesephist.comWithout that community at the top of the funnel, we’d be just like everywhere else, a top-heavy ecosystem wondering why the brand-new makerspaces and meeting rooms sit empty, while the business competitions are won by MBAs who go on to spend the rest of their twenties at McKinsey.
Where Cal founders come from | thesephist.comTowards the tail distributions on both traits, unexpected tail events start to happen that dominate everything else
Betting on Unknown Unknowns - by Alexandr WangWhen Moore first formulated it in 1965, we did not know all of the innovations or feats of engineering that would have enabled this trend for the next 55 years. But, Moore was right, and every year we have continued to make advancements in research to continue to drive meaningful improvements.
Betting on Unknown Unknowns - by Alexandr Wangthey are betting on “unknown unknowns” which will meaningfully change the game.
Betting on Unknown Unknowns - by Alexandr WangAs the pace of technology has accelerated, the rate of surprises has as well. It’s not hard to find examples of experts’ predictions which always end up being way off. It’s not because the experts are not knowledgeable—in fact it might be because they are too knowledgeable.
Betting on Unknown Unknowns - by Alexandr WangSaving a person's life feels great, and it would probably feel just about as good to save one life as it would feel to save the world.
On caringThe problem is, my brain is simply incapable of taking the amount of caring I feel for a single person and scaling it up by a billion times.
On caringOne very scary thing to me is Scale becoming a credential rather than a cult.
Hire people who give a shit. - by Alexandr WangIf you want to be a writer, you keep writing. Even when you’re unpublished, even when you’ve gotten rejection after rejection after rejection.
Why I Write: On Soul Debt and Min Jin Leewhen you write from a place that is real and genuine to you, you will never have writer’s block. You might be an incredibly slow writer, the way she deemed herself to be (she took around ten years to write her first novel), but you’ll always have something to say. Write from a place where you feel visceral, intense emotions. That’s your motor. Keep writing from there.
Why I Write: On Soul Debt and Min Jin LeeBeing independent in a healthy way doesn’t require avoidance, it requires self-differentiation: staying yourself in the presence of people who are important to you.
how can we be the right kind of together? - by AvaNo one can help you self-actualize except you.
how can we be the right kind of together? - by AvaIn brief, people with self-respect exhibit a certain toughness, a kind of moral nerve; they display what was once called character, a quality which, although approved in the abstract, sometimes loses ground to other, more instantly negotiable virtues.
On Self-Respect: Joan Didion’s 1961 Essay from the Pages of Vogue | Vogueself-deception remains the most difficult deception.
On Self-Respect: Joan Didion’s 1961 Essay from the Pages of Vogue | VogueI had somehow thought myself a kind of academic Raskolnikov, curiously exempt from the cause-effect relationships that hampered others.
On Self-Respect: Joan Didion’s 1961 Essay from the Pages of Vogue | VogueWe have to be willing to challenge the conventional wisdom, which in this case means asking not how to improve grades but how to jettison them once and for all.
The Case Against Grades (##) - Alfie KohnGrades don’t prepare children for the “real world” — unless one has in mind a world where interest in learning and quality of thinking are unimportant.
The Case Against Grades (##) - Alfie KohnThey end up learning more once they aren’t feeling the pressure” from grades.
The Case Against Grades (##) - Alfie KohnTheir writing improves more quickly and the things they learn stay with them longer. I’ve had lots of kids tell me it’s changed their attitude about coming to school.
The Case Against Grades (##) - Alfie KohnRobbins says he almost always accepts students’ suggestions because they typically pick the same grade that he would have.
The Case Against Grades (##) - Alfie KohnThe claim here is that we should do unpleasant and unnecessary things to children now in order to prepare them for the fact that just such things will be done to them later. This justification is exactly as absurd as it sounds, yet it continues to drive education policy.
The Case Against Grades (##) - Alfie KohnWhen comments and grades coexist, the comments are written to justify the grade
The Case Against Grades (##) - Alfie KohnIf you’re sorting students into four or five piles, you’re still grading them.
The Case Against Grades (##) - Alfie KohnBut grading for learning is, to paraphrase a 1960’s-era slogan, rather like bombing for peace. Rating and ranking students (and their efforts to figure things out) is inherently counterproductive.
The Case Against Grades (##) - Alfie KohnThey offer a way to thoughtfully gather a variety of meaningful examples of learning for the students to review.
The Case Against Grades (##) - Alfie KohnOnce we’re compelled to focus only on what can be reduced to numbers, such as how many grammatical errors are present in a composition or how many mathematical algorithms have been committed to memory, thinking has been severely compromised.
The Case Against Grades (##) - Alfie Kohnthe more students are led to focus on how well they’re doing, the less engaged they tend to be with what they’re doing.
The Case Against Grades (##) - Alfie KohnIf nourishing their desire to learn is a primary goal for us, then grading is problematic by its very nature.
The Case Against Grades (##) - Alfie KohnThey’ll choose a shorter book, or a project on a familiar topic, in order to minimize the chance of doing poorly — not because they’re “unmotivated” but because they’re rational.
The Case Against Grades (##) - Alfie Kohnevery study that has ever investigated the impact on intrinsic motivation of receiving grades (or instructions that emphasize the importance of getting good grades) has found a negative effect.
The Case Against Grades (##) - Alfie KohnCollecting information doesn’t require tests, and sharing that information doesn’t require grades. In fact, students would be a lot better off without either of these relics from a less enlightened age.
The Case Against Grades (##) - Alfie KohnIn other words, what we should care about when it comes to A.I. in the real world is not explanation. It is validation.
Software vendors are pushing explainable A.I., but those explanations may be meaningless | FortuneWhat’s more, they found that in real-world settings, most people using the algorithms had no way of resolving such differences and often, as Ghassemi, Oakden-Rayner, and Beamer, suggested, simply picked the explanation that best conformed to their pre-existing ideas.
Software vendors are pushing explainable A.I., but those explanations may be meaningless | FortuneIf the feature strikes the doctor as counterintuitive, what should the doctor do: conclude the algorithm is wrong, or conclude it has discovered some clinically-significant clue previously unknown to medicine? Either is possible.
Software vendors are pushing explainable A.I., but those explanations may be meaningless | Fortunethey can produce broad descriptions of how the AI system works in a general sense but, for individual decisions, the explanations are unreliable or, in some instances, only offer superficial levels of explanation.
Software vendors are pushing explainable A.I., but those explanations may be meaningless | FortuneEveryone who is serious in the field knows that most of today’s explainable A.I. is nonsense
Software vendors are pushing explainable A.I., but those explanations may be meaningless | FortuneWait until you meet someone for whom telling them about opportunities actively hurts them, because you’ve just created another knife they feel pressured to cut themselves with.
PainWhen a theory breaks, you learn how to build it better
How to break a theory | symmetry magazineWe value the time of others because we know that is finite. Life has meaning because it is finite.
nothing stops