Tasha

@shawkisukkar This is why I write in a physical notebook next to my laptop at work

I watched the first ten minutes of the Bill Gurley and Garry Tan interview. I had to turn it off, but what I realized is that the highest-leverage thing you can do is to use LLMs very minimally. If Gurley is saying something like If you don’t use OpenClaw, you’re going to get left behind, and agreeing with Tan; there’s something very wrong going on in the current consensus. You can’t outsource thinking, and you can only really think through the hard process. The thinking that happens through the details is what determines the winners and losers. This is why startup founders don't hire a consulting agency to build their product. It’s that process, and the obsession it requires, that makes people think of great ideas. This is why being forced to raise very little money is often a blessing for founders. There's no such thing as learning to think from first principles. You just work through every little detail as hard as you can while knowing when to zoom out. There haven't been any impressive AI features except for the chat interface decision. The entire lines of products and how people choose to integrate LLMs, build on top of them, etc., all show a lack of intention. There's a lack of obsession and care because people don't think hard anymore. Linear built a world-class contrarian product at one point, and last week they released a below average blog post about their "new vision". You can’t understand the great books by reading books about them, it’s the same with work. I don't want LLMs to summerize docs, or plan a technical implemention anymore. I want to feel every little detail, work hard for every button, feel the serendipity of every unimportant thing.

@CheckCanopy @trishaepan wow congrats

It’s hard to believe it’s already been 100 days since I received my Neuralink N1 implant. Looking back, the whole journey feels like science fiction that somehow became my everyday reality. The surgery on Day 0 was surprisingly easy. A quick general anaesthetic, a small incision, and the robotic system did the rest — precisely placing the 1,024 ultra-thin threads into my motor cortex. I woke up alert and in good spirits and went home the next afternoon. By Day 3 I was feeling a lot better, and by Day 7 the little scar was already starting to fade. Recovery was genuinely minimal; I felt sharper and more positive than I had been in years after the BCI was turned on. The real fun started in Week 2 when we paired the implant with my brand-new Apple MacBook (my very first Mac). The @neuralink engineers walked me through calibration sessions, and within a couple of minutes I was moving the cursor just by thinking. At first it felt like trying to remember a dream, but by Week 3 it was second nature. Scrolling, clicking, typing — all mind-controlled. The Mac integration was buttery smooth; I went from total Mac newbie to power-user faster than I ever expected. By Day 80 I was ready for the big leagues. That’s when I fired up @Warcraft of Warcraft for the first time with pure thought control. The first raid felt clunky, but once my brain and the BCI synced, it was pure magic. I’m now raiding, and exploring Azeroth hands-free at full speed — no mouse, no keyboard, just intention. It’s honestly brilliant. The freedom is addictive. The social-media side has been just as surprising. Every update I’ve shared has been met with genuine excitement rather than scepticism. Thousands of messages from people with disabilities, gamers, students, and scientists — all asking real questions about the tech and what it could mean for the future. The positivity has been overwhelming and incredibly motivating. 100 days in and I already can’t imagine life without it. The N1 didn’t just give me a new way to use a computer — it gave me a new way to live. Can’t wait to see what the next 100 days bring. Thank you all so much for your support and I will keep you all updated as we continue this journey together.

@NaderLikeLadder @DanielLurie Do you plan on having many tickets to pay?

Hey @DanielLurie 👋 my openclaw can’t pay my parking tickets because of the CAPTCHA. If I built an SFMTA CLI, would you let us use it?

I wish cursor had a desmos integration

@keerthanpg Yeah I’ve always wondered how research startups that raise money to work on toy problems like better learning efficiency are able to prove their construct validity. Like how are you certain your findings will hold up on a much bigger model?

Big model robotics is qualitatively than small model robotics. Lots of insights and “truths” discovered at small scale stop holding up once you scale up the model - much like in LLMs.

@BerkeleyML @sherjilozair

who are the best people to talk to about physics + ml?

@deedydas Didn't see @softmaxresearch on the list

The Ultimate List of Artificial Intelligence "Neolabs". A Neolab is a pre-revenue scale startup working on long-term AI breakthroughs. Here's all 50 of them.

@owl_posting Is the point of this to say that humanity doesn’t make progress

our apartment cat—a two-year old girl named Cinnamon—recently had surgery to fully remove her nonfunctional eyes and sew up the holes. despite how horrifying this sounds, she seems much happier! more curious about her surroundings, less aggressive, and now apparently capable of precognition. two days ago, after arriving back to new york, she indulged me in a game of chess before i went to bed. she has a great memory, and is usually able to play decently if i announce each of my moves in standard Algebraic notation. but in last nights game, she told me there was no need to announce anything at all; she could understand the state of the game just fine. i was stunned. i asked her if she could somehow see. she told me no, i am still blind, but that the removal of eyes branded onto me an instinctual understanding of how the tides of reality will bend in the next few seconds. i know who will die, who will be born, and what all chess moves made in that timeframe will be, including your Qh5 move. it is hard to explain, she purred, i cannot see any of this. i cannot see the white or black of the board, nor anything else, but information arrives fully crystallized, and i am able to act upon it. i told her that that sounded overwhelming. she yawned and said, well, its a bit of a relief actually. i used to spend all day worried about every little thing. the climate, wars, political polarization, and on and on and on. so stressful! but now i can just see the truth, every single possible truth. every aneurysm, every revolution, every turn of the gear in the system. and i must admit: its all banal, horribly banal. i could imagine feeling a sense of responsibility about it all if i were not, say, a cat. but i am a cat, and so i am happy to listlessly stare at it all and think to myself, well, that’s that. that’s the grand secret. guess ill go take a nap. i briefly wonder if i would feel a sense of responsibility if i had the same abilities, and Cinnamon takes the moment to checkmate me. i ask if she thinks that this newfound ability has made her a better chess player. she laughs and says no, i just got lucky here. there is another chess game being played in Caracas, Venezuela right now that happened to mirror our exact game almost perfectly, and i used their game to help avoid dumb mistakes. i immediately accuse her of cheating. meow, she responds back

@a_karvonen @_catwu If you’re in a tmux window, ^B[ to scroll and then ^B[q solves this to refresh claude

@_catwu I have poor reliability when using Claude code cli in VS Code's remote ssh extension. I usually have to reload the window because Claude code has frozen every three prompts or so.

We'd love to hear your Claude Code feedback!

@gowthami_s I wish

Don't hire based on what they know or have done in the past! Hire based on what they are capable of and how much they love the problem!

@relh_net @slow_developer How do you stay sharp as a developer with this? DRY, type safety, separation of responsibilities, avoid stateful code, etc. but is it really just “spend 30 years writing code and apply unreasonable amounts of perfectionism”?

@slow_developer It's like we've all become managers or PMs without realizing or even wanting to, it's just so much faster to have 4-5 copies of the repo checked out locally on different branches and to jump between each with instructions

Anthropic engineer: "the last month was my first month as an engineer that i didn't open an IDE at all, and opus 4.5 wrote around 200 PRs, every single line" you can already sense what's coming next year. more and more devs are becoming the ones directing the work, not the ones typing it out

@KrishivThakuria aren’t you in high school cool!

Will be teaching this course at MIT EECS in January If you're at MIT and thinking about enrolling in an IAP course, S189 will be a good time!

@yacinelearning @softmaxresearch It’s a difficult problem bc agents have to maintain beliefs over hidden states rather than condition on true state, which explodes the state space and makes learning less sample efficient

@TashaPais @softmaxresearch no way so cool how’s it going so far?

if we were to extrapolate the pace of RL research coming out of open ai we would have solved WoW by now what are we even doing man

@dwarkesh_sp @langofmind has an excellent YouTube channel and was my fav prof

Looking for a neuroscientist to interview on my podcast. Keen for someone who can draw ML analogies for how the brain works (what's the architecture & loss/reward function of different parts, why can we generalize so well, how important is the particular hardware, etc).

Are any of the Ada paper authors at NeurIPS? If so, would love to talk abt extensions to this paper from 3 years ago arxiv.org/abs/2301.07608 @jkbr_ai @katebaumli @mmmbchang @edwardfhughes

@huskydogewoof 👋 would love to talk abt games and memory

If you are interested in world modeling things like game simulations, memory, and data, I would love to chat from Dec 3 to 7 at San Diego! Recently I have also been very into deep equilibrium / recursive / loop models, and I am happy to discuss about those too. I will probably be hanging out next to my awesome roommate most of the time, adding the world modeling corner to the expertise he listed here 🥸 so dropping by is basically a two in one combo.

@callmeMaharani read at a bar alone

what do single people who aren’t actively dating and spend most of their time alone even do on weekends and every night ? Asking for a friend

@eigenron Will talk abt game envs for models

i'm going to #NeurIPS 2025! lmk if we should meet and chat about RL on diffusion, world models, gaming environments for models, etc.! it could even be physics or movies lol.

@wandb !!!

🤫 Something's been brewing in stealth. Our SDK team's side project, codenamed W&B LEET, is being unleashed. We are releasing a full Terminal UI (TUI) for live, interactive W&B monitoring right in your terminal. No browser, no internet, no problem.