babak damavandi

@MillionInt bitter is taken, would you like the Rancid Lesson™

Low bias high variance methods are usually not very compute efficient but they are scalable High bias low variance methods have higher efficiency but hit limits of scaling quickly One of those sides has had better time historically

we brought live ai to Ray-Ban Meta glasses a year ago. Starting today, you can use live ai on your phone too. Enjoy!

@thsottiaux oh and if you could please don't drain my battery in like 7 minutes with electron that'd be so cool

Become the interface to everything on the computer. Nobody really wants to click endlessly or to figure out where the UI buried what they care about. Make /goal a first class citizen. Let us use it interchangeably for projects. "Make sure I do cardio twice a week" should totally work as a goal. Agents control panel that feels like I'm talking to a team of humans, not running scripts in random directories. Fix the whole working directory context. Programmers get it, normal users don't, and they shouldn't have to. I can go on. I love codex. Keep executing!

Now that the Codex app is close to being the super app. What should the super duper app do?

kids, by May 2026 we had AI agents that browsed the human web for us, but refused to click I'm not a robot. It was a confusing time.

@_arohan_ is it a shampoo shop though

It’s core automation. We are building the world’s most automated ai lab. I repeat, we are not a car. We are not car auto, I repeat we are not car auto or core automative.

@gvanrossum As someone who's lived in both countries for decades, respectfully, I don't think you have any idea what you're talking about.

From a NYT article about Ayatollah Khamenei: “He was at once revered by followers as a representative of God, and as the commander in chief of the armed forces, with the definitive word in all key state matters.” Sounds a lot like Trump, except for the God thing.

@NYCMayor We're in New York because we fled this regime. You really don't have to comment on topics you have no idea about.

Today’s military strikes on Iran — carried out by the United States and Israel — mark a catastrophic escalation in an illegal war of aggression. Bombing cities. Killing civilians. Opening a new theater of war.  Americans do not want this. They do not want another war in pursuit of regime change. They want relief from the affordability crisis. They want peace. I am focused on making sure that every New Yorker is safe. I have been in contact with our Police Commissioner and emergency management officials. We are taking proactive steps, including increasing coordination across agencies and enhancing patrols of sensitive locations out of an abundance of caution. Additionally, I want to speak directly to Iranian New Yorkers: you are part of the fabric of this city — you are our neighbors, small business owners, students, artists, workers, and community leaders. You will be safe here.

@signulll I basically haven't slept since mid Dec. We're past the event horizon

this is going to sound dramatic as hell but the world you knew is pretty much over. very recently i think we crossed a one way bridge as a species & most people on earth haven’t realized that fact yet. imho there are now non trivial odds the economy gets drastically disrupted in the next 24 - 36 months via a giant supply shock followed by a demand shock.

Essays are AI generated, their summaries are AI generated. The last remaining artifact of original thought will be the OG 140 character tweet

Try explaining agentic coding or spacex transitively owning twitter to your pre-pandemic self

We are the bootloader for LLMs. LLMs will be the bootloader for AGI. Doesn't matter how many tricks remain, recursive improvement is the endgame for LLMs

@paultoo we're building something ai wants, to spend the money on what ai feeds tell us we want

make something ai wants

‘Not piling on, but’ proceeds to pile on.

“Context left until auto-compact” a horror story in 5 words

@JacksonKernion I’m invested in this one, mostly because I’d love to see us use LLMs/AI to understand and ultimately rewire our own nervous systems. It could be an AlphaFold type moment for psychotherapy.

2. LLM psychology: understanding the unique aspects of machine intelligence, developing evals for key cognitive traits, explaining the contrast with human intelligence.

I'm trying to figure out what to care about next. I joined Anthropic 4+ years ago, motivated by the dream of building AGI. I was convinced from studying philosophy of mind that we're approaching sufficient scale and that anything that can be learned can be learned in an RL env.

@alxfazio It’s annoying. I just tell it that another agent needs to work on X and to dump instructions and all the required context into a .md file. Works, but very clunky.

claude code needs conversation branching badly. a lot of the time, something comes up mid session that i need to dig into, and i’m stuck choosing between two annoying options. either i spin up a separate conversation and manually carry over the context, or i finish what i’m doing first and only then scroll back through the chat to pick up that side thread

@Strippin this has 6.7k likes and you still think we're not in a simulation

My 7 year old daughter just asked me why adults hate counting to 8 Me: 'why?' Her: 'cos 6 7' Then she moved her hands like she was weighing fruit at a market???

@signulll We took it for granted.

if they ever release old google TGIF’s of larry & sergey, the world would know how fucking hilarious these two dudes were up there every damn friday. it was like watching a live comedy show all while knocking back few beers at charlie’s. what a time that was. i don’t think many ppl grasped how unique it was to be a google employee back then.

@Noahpinion Compared to SF? No. Compared to Tokyo? Yes.

New York is great but its street fashion is consistently bad