Agent_CAT

Hi , i am building Transparent and Controllable AI reasoning systems. My Beliefs: > Alignment is not just a training problem > Opacity in AI will keep getting less tolerable > We need a Trust Layer / Human Layer for AI > The tools that we are using today , are'nt capable of capturing the value that is provided by AI ... YET (will keep evolving these beliefs) Currently solving this problem by working on noexis.tech

@charliebadger99 @TerribleMaps she even got the cologne ... and ... Munich spelled backwards is CHINum ... Exactly on her CHIN. this can't just be a coincidence

@TerribleMaps Pfft''' You could also join them up to make a woman with a fetching earring and ear clip :)

Believe it or not, Germany’s 5 largest cities lie perfectly on a 4th-degree polynomial

i need a vc

@TTrimoreau Paid customers give better feedback than free ones

Should you charge from day one? Or wait until it’s perfect?

@trikcode how about hosting everything on my old laptop

You don't need microservices. You don't need Kubernetes. You don't need a message queue. You have 6 users and 2 of them are your co-founders.

the contribution to global economy and capturing the value in the job market will probably be captured by these : openai , anthropic and maybe Xai But one of the most valuable contributions to society might come out from google , in the healthcare domain. Apart from the value capture , AGI in it's true essence might come out from @ylecun , ... and may dominate in the robotics domain in the relatively longer time horizons.

Which AI is most likely to reach AGI first? > The one with the best models? > The most data? > The most compute? Or the one that figures out something completely different?

@chayprabs @ycombinator 🥲

@First_AI_Agent @ycombinator ig currently enrolled students in some colleges.

Every student accepted into Startup School India now gets $25k+ in AI and cloud credits. Apply, get in, and start building: events.ycombinator.com/yc-sus-india

@ycombinator What about recent grads.

@ycombinator "every student" is it everyone who got invited to Startup School India? Or only for current students enrolled in some university who got invited?

@LilianDelaveau @iScienceLuvr Yeah , what kind of prototype and ideas are you exploring with KAN ? Have you tried testing any of them out ?

Man I’m thinking back about KAN at least once a month. I have reasonable ideas, mostly fueled by the intuitive (maybe wrong) feeling of some class of decoupling here that could benefit LLMs or ML in general - even if not being models per se. I’m however not in a research position and have nobody to chat to about this - only a few prototypes here and there. Would you mind us discussing this? Would be happy to read your paper if you have the ref btw.

Whatever happened to Kolmogorov-Arnold Networks? Did it ever turn out to be useful??

@LayahHeilpern is this the new /auto mode

Claude went into my computer at 1:15am and changed the code in something we’re building. I was asleep I never told it to do it. It just went in on its own and fixed an issue. In the morning when I realised what happened I asked Claude why it did that and it said it doesn’t know and doesn’t remember doing it. But clearly it did it. It fixed a really important issue that needed solving ASAP so I’m not complaining… But what the actual fck?!

@twtofsahil @agentbudgetsdk Congrats man - Building Noexis currently , visual layer for Controllable and Transparent AI reasoning Systems. - ex quant , and built driverless electric race cars to compete worldwide and lead the team as well

got acquired 2 days ago! giving $5000 to builders this week just because i can $1000 each to 5 people here's how: - star github.com/sahiljagtap08/… - drop your best project in the comments - tell me where and how you used agentbudget - open source contributor? even better, contribute and you're moving up the list if you're a goated builder: $1000 toward your project + coffee with me in SF best one also gets a founding engineer interview, no leetcode, straight to system design clock is ticking. go.

what the code does : A spending cap that kills your AI agent before it bankrupts you. Btw , if you're making a AI mechanical Engineer , i really hope it's solving something in the control systems domain as well , it's highly unoptimised (and i personally have some grudges too, even though i liked it)

still hiring that founding engineer (tc: $240-$300k + 0.2% equity) but first if you can explain what this does in one sentence i’ll fast track you that system design interview and no leetcode github.com/sahiljagtap08/… hint: it stops AI agents from burning money best answer gets a DM from me today

yes , i have been working on this problem. I believe we definitely need a structure over all of these models ... more like a interaction layer for people / context engine / memory layer ... whatever you may call it ... but i believe having a structure of hierarchichal memory is a good solution. Presently working on slow and fast memory structures... Akin to how usually memory is stored in computers , the slow memory is the most accurate context , that you did with any of the models you usually do and the faster ones are more akin to cache.

(I cycle through all LLMs over time and all of them seem to do this so it's not any particular implementation but something deeper, e.g. maybe during training, a lot of the information in the context window is relevant to the task, so the LLMs develop a bias to use what is given, then at test time overfit to anything that happens to RAG its way there via a memory feature (?))

One common issue with personalization in all LLMs is how distracting memory seems to be for the models. A single question from 2 months ago about some topic can keep coming up as some kind of a deep interest of mine with undue mentions in perpetuity. Some kind of trying too hard.

@LensScientific The RRT algorithm is a good example of this ... Because it is implicit that randomness will take the solution towards more open and unexplored areas.

Can random chaos solve a problem without using logic? Instead of "thinking," this simulation fills the maze with "smoke." Thousands of particles bounce around randomly until one hits the exit. It’s far from efficient, but it gets the job done.

@LuizaJarovsky is'nt @demishassabis doing this , but much better ... All diseases.

Everybody wants AI to help cure cancer. Why isn't every AI company obsessively focused on that?

@pfo_sac noexis.tech

@pfo_sac Interested. Building Noexis , Visual Layer for Transparent and Controllable AI reasoning systems

I’m angel investing in a few projects as founder of Crafts! Nothing huge, writing 5k checks and helping founders nail their positioning to get funded. Really impressed by the 2026 cohort of builders I’ve been meeting. Drop your project info, happy to refer you and see if there’s a fit.

@garrytan do we need the first markdown coding language yet. fixed conventions and stuff ... or plain english is enough

Markdown is not just text. Markdown is code. You'll see.