rohit

@yacineMTB hotter take: knowing how to properly install python/packages [pip/uv] is/was higher value

hot take: if you are a programmer you should be able to invert a binary tree from memory, AI or not. It's ridiculously easy and if you can't do it, you should not have a computer science degree

@vasuman codex 5.5 extra high, fast mode

@benhylak @cremieuxrecueil

@cremieuxrecueil yep. as a founder, i've had to mostly stop calling out fraud because i get in trouble.

It's always very funny when a founder commits fraud and the VCs who backed them get angry at the person who found the fraud rather than the founder who defrauded *them*.

@FangYi11101 Boot up cluely 🥲

interview candidate tmr wants to do the technical round in C++ fml

How much did the thinking machines model cost to train?

for a social company, why does metas consumer ai deployments seem so poor? It’s almost on par with copilot

the quality is insane

@staysaasy @olvrgln Does starting at N on every dimension open up new opportunities, or just create mediocrity

It’s 2018 and your coworker just sent you a 400 line pull request. You get a cup of coffee and sit down to review it. It’s beautiful. Elegant micro-refactors. Crispy method names. You catch a few things, but that’s ok. It’s part of the dance. They didn’t consider extensibility on part of their API. Here’s a comment buddy. They respond in an hour saying they think we should do one piece differently than your comment. Hey let’s jump into a room and figure it out. We can’t just agree to disagree, this code is too important. The PR merges and goes to prod. You feel a shared sense of ownership and accomplishment. That night you go to sleep and dream of that code. You can still see the shapes of it on the backs of your eyelids, your IDE syntax highlighting sparking neurons in your reptile brain. You go to work the next day ready to go. You understand the system. N is your foundation. Time to build n+1.

@justinskycak Bro is roasting the league & valorant community in one tweet

Avoid games where the prize is validation from people you do not respect.

@xuanalogue agreed. while it’s an exciting idea, it’s so slow (and this was the small model 😭)

It's refreshing to see an AI lab go in a different direction! But also, as someone who's obsessed over getting my inverse planning algorithms down to milliseconds of latency, I feel like I'm being gaslit when they caption those 1.5s to 3s delays as "instant reactions" 😵‍💫

My reply to someone considering starting a video game company: The distribution of possible rewards for starting a video game company are generally not very good today. The market is well served, and gaining a foothold requires strong execution on both business and product issues, along with a substantial amount of luck. Plan to burn through seven figures with a not-great chance of making it back. If you do go for it, some bits of advice: Identify your customers clearly before you start. Not just a broad community, but specific people, and imagine them as you make decisions. Initially, build the smallest, most concise game you can imagine anyone paying for. It will still take much longer than you expect. Once something exists, hill-climb the value. Hopefully you will have some elements that clearly bring joy to people, which you can magnify. There will inevitably be tons of things that people find confusing, frustrating, or just boring that you will need to fix.

$GME looking more and more like $HLF

codex + gpt5.5 extra high fixed it, turns out you need the worlds most powerful intelligence model to get windows to play audio correctly, ty @sama

I used to clown on linux users for driver issues, turns out windows is strictly worse these days :/

@ericcouu I think you mean alcatraz

hot take: golden gate park should turn into the world's largest datacenter

@BenjDicken @grok wat dis, explain using emojis

"I think I'll study up on database query planners" Every query planner paper ever:

the ai talking over the humans part of the demo feel a bit weird, not as natural as Moshi, but still impressive that it retains context or multiple speakers.

thinking machines enters the voice model game

@thinkymachines 12/10 acting 🤣

Tessa's quality of life has improved a lot with some nagging.

People talk, listen, watch, think, and collaborate at the same time, in real time. We've designed an AI that works with people the same way. We share our approach, early results, and a quick look at our model in action. thinkingmachines.ai/blog/interacti…

That makes sense. I rank moshi's speech flow as #1, but I find that sesame's voice models are able to almost close the gap via high quality actors/voices/empathy/phrasing. Hopefully the breakthrough comes soon, but in the meantime there seems to be a high ceiling for cascade models.

Because that's what people are using right now. A big advantage is the flexibility you get from being able to just switch the LLM, TTS or STT backend whenever a new model is released, taking advantage of several providers jointly. Also the cascaded latency we're getting today is much better than it used to, reducing the gap with speech2speech. Yet, turn-taking is a curse and full-duplex is inevitably the horizon, but we cannot expect users to give up any intelligence for that. So this will take some new scientific breakthroughs to combine the naturalness of Moshi with the abilities of the best cascaded systems.

great talk by @neilzegh on where voice models are today youtu.be/P_RI1kCkRbo?si…