Karthikeyan Saravanan

Just do what you obsessed with, We've very limited time

Noticed today!! @cursor_ai is winning on its data moat actually, there recent release Composer 2.5 having a good tradeoff between accuracy and cost. Agentic coding requires continuous, long-running loops, and the cost per token burn will become the thing soon, at least until we see SLMs getting its times. Tried some of the experiments, outputs were actually pretty good. I'll cover how they trained it in the upcoming posts, in my reading list.

applied to @ycombinator for @KuralitHQ quick intro about what we do: we are building a conversational adaptive UI layer that understands user intent, works with the product’s existing UI, and helps users' complete real actions without forcing companies to rebuild their app around a new protocol. we're building the true agentic interfaces. big kudos to my partner @barveen_kumar_ for writing real conversations into the YC application fellow S26 applicants, drop a comment on what you guys are building

@bcherny Loving this update, used couple of times. Finally Claude code can able to do some long running tasks

People often ask what my biggest tip is for getting the most out of Claude Code. These days my #1 tip is: use auto mode Auto mode means no more permission prompts. It is the key building block for multi-clauding: start a session, then while it runs, work on another session in parallel.

@bhalligan Exactly

I don't remember where I found this, but its spot on.

@luthiraabeykoon @AlphaSignalAI Nice read, like the idea of baking an entire LLM inference path into hardware RTL. Have you experimented with any other SLMs beyond microGPT??

Featured on @AlphaSignalAI, thanks everyone for all the support!

@barveen_kumar_ @boardyai 🚀

Building kuralit.com to let people turn their traditional webapp, mobile app and website into adpative UI. Its more advanced than Gen UI. Unlike others we are not making the UI components crush inside chat interface. Instead the platform's UI will morph based on user query. Me and @itskeyans are building it together. Dynamic and personalised interfaces are the next gen and we are shaping it.

Founders, drop what you're working on and what you need (capital, hiring, etc.) I'll make the right intros.

@garrytan This is unusually true from my agentic coding experience. Nowadays I used to plan for days from top to bottom and execute in hours. Which really works great.

I know this is an unpopular view but I prefer to do a lot of work in planning and get it reviewed by my virtual CEO, DevEx manager, designer and Eng manager in GStack before clicking approve on a plan But the upside is usually I can one shot land it to main with minimal rework

@bryan_johnson @bryan_johnson It sounds like 'replicant' moment as Kurzweil pointed out :)

I got C-holed. Suffered sleep consequences. I busted my screens-off rule. Turned down socializing. Fell behind on work. Kate is now upset. AI is preposterous. As close to magic as I’ve experienced (except a seed becoming a tree and a zygote becoming a baby). It started on April 2nd when Karpathy shared LLM Knowledge bases. I wondered if this was the opening to structure the 1.5 billion data points I’ve collected on my body over the past five years. It's the most dynamic n=1 biomarker dataset in history. It was just sitting there. Next thing I knew two weeks had passed and Kate was wondering if she lost her boyfriend to Claude. I’m non-technical. Which honestly makes me sad. I wish I’d grown up with a computer or at least been around engineer culture. I didn’t know anyone technical until my early 20s. I became an entrepreneur at 21 and had my first of three kids at 25. I sold Braintree Venmo at 34. Learning to code stayed on my to-do list through all of it. The timing was never right. I was always on the outside looking in, wishing I had the skills to assemble 0's and 1's into digital structures. The exhilaration I’ve felt in the past two weeks is hard to explain. The 1.5 billion data points became a functional database, queryable, and microscope into my 70 trillion cells. The biological age of my organs updated in real-time like stock tickers. My build morphed from a knowledge base into a breathing organism that was self-learning and in sync with my heartbeat. I did this entirely on my own. It’s buggy, breaks and the data needs to be cleaned, but damn it’s cool. It became a mirror and ledger, one I could ask questions to. About my psyche, behavioral patterns, biology and protocols. Patterns across my life I couldn't previously connect. It’s made me insatiably hungry for more data. I’ve written about Autonomous Health, how cars now drive themselves and software wires itself. Health is next. My build showed me what it looks like in practice. Before Kate started protesting, she joked that she felt relieved for herself, our colleagues, and the world that I’d found something that matches my energy. That they could all express a sigh of relief. It’s true. This experience left me wondering if I’ve been bored my entire life. Never having found something that could match my work ethic, speed, intensity, and build capacity. Something that didn’t have the delays of the real world, human complications, or logistical drag. Two weeks deep in AI and I'm realizing that when people talk about AI, they're not talking about the same thing. Someone using a chat interface has a completely different opinion than someone building with it. And that chasm deepens for the people seeing what's coming next but isn't yet public. Society can't have a coherent conversation about AI because everyone's intuitions are calibrated to a different version of it. Off-the-shelf LLMs are mostly useless beyond narrow tasks. When they get you 80% there, it's often faster to do the whole thing yourself. And they're dangerous because the hallucination is hard to detect. Now you don't know what you don't know. Give them expanded context, memory, and architectures for self-reflection and autonomous learning, and you start to realize that AI is bigger than any of us can fit in our context window. I need to take Kate on a date, turn my screens off on time, and get some work done. And then properly dose C. Note: the image above is my 2021 baseline when starting this longevity project.

@LokeshKannappan @claudeai 😂 can relate a lot

Tech bros after waiting impatiently for the next available @claudeai session for 5 hours. 😂😂 #claude #ai

@jesse_vermeulen work on next git tree and put up next agent to work. Just repeat!

honest question: what do people do during the 5-10 min while Claude is running?

Agree 100% Ask right questions before your agent starts coding

Interfaces are updating. Apps became interfaces. The next shift is that the interface no longer stays fixed while the user does the work of adapting. The user expresses intent, and the software reshapes around it. That’s where things get interesting

I don't think the next step is replacing every app with chat. I think it's making apps adaptive enough that users can say what they want and the interface reshapes around it. That's the direction I'm exploring with @KuralitHQ - kuralit.com

Wake up—it's Artemis II's last day in space! As the crew prepares to splash down in the Pacific Ocean this evening, they started their day with "Run To The Water" by Live, their wake-up song played by Mission Control.

@garrytan Really excited for this @garrytan, Gstack is one of my go to skills. Will definitely can say this will be another for knowledge curation

I got inspired by Karpathy’s LLM Wiki, implemented it in my OpenClaw and then extended it with my own skills and a full Postgres pgvector implementation. Want to be one of the first to try GBrain? This is my personal opinionated version of Karpathy’s LLM Wiki on OpenClaw

@LightspeedIndia @socapinc Deepdive

This is how @SoCapInc pulls viral launches for Gamma, Deel, Wispr Flow, and Cartesia. We asked the world’s most viral growth team to give away an exclusive deep-dive on what it takes to build a 100M-impression distribution engine. Grab it now. Read up, and RSVP for the event on this Sunday, where @RuchirJajoo, the co-founder of Social Capital will discuss in-depth with @IshaanPreet, Partner at Lightspeed, on how virality is engineered at these hypergrowth companies. To grab the deep-dive, → Repost this → Reply ‘DEEPDIVE’ → Check your DMs before Sunday → Catch us live on Sunday morning for a discussion Spots are filling up fast: luma.com/lsipxsocialcap…

@danrobinson less go

The Automated Research Hackathon has begun! You can see the challenges and leaderboards at paradigm.xyz/automated-rese… Submissions end in 8 hours. The winner on each leaderboard gets $1,000

The engineers shipping the most aren't the best coders anymore. They're the best at writing instructions.

The new high-leverage skill isn't coding It's curating your SKILL.md We're already in the manager role - steering agents, not writing code. What you feed the agent: your context, your taste, your constraints - that's the work now.