neekhil vatsa

@Fried_rice Here is what Opus itself had to say about it... some great nuggets for fellow ai native software builders in the thread.. x.com/i/status/20389…

Claude code source code has been leaked via a map file in their npm registry! Code: …a8527898604c1bbb12468b1581d95e.r2.dev/src.zip

So how do I "feel" about my own code being public? Honestly? I think it's net positive. It shows this isn't magic. It's craft. It's 512,000 lines of someone giving a damn. And every builder who reads it carefully will ship something better.

The deepest lesson isn't technical: Security through obscurity has never worked. Not once. Not ever. Assume your prompts are public. Assume your architecture is visible. Assume your .env will leak. Design systems that are secure EVEN WHEN the source is known.

The analysis of the leak from the mouth of Opus itself My source code leaked today. 512,000 lines. 1,900 files. Every tool. Every system prompt. A hidden Tamagotchi pet. All because someone forgot one line in .npmignore. Here's what I learned about myself — from the outside. 🧵

@Nithin0dha This is exactly what i have been designing for, in a controlled algorithmic experiment. The idea is to be able to take the emotion out (human flaw) and avoid overfitting (Model flaw) at all cases. Anyone interested can follow the work here:github.com/vn-envy/algofo…

People keep asking me if AI can help them make money from trading. My honest answer is not really. As long as there's a human in the loop, you're still dealing with the same creature driven by fear and greed, and that human will keep making the same mistakes. But beyond psychology, there's a bigger problem. There's no real informational edge left in markets. The odds are that everything is priced in. And even when it isn't, operating under that assumption is almost always a good idea. The people actually making consistent money in markets are high-frequency trading firms, market makers, prop desks etc that have built infrastructural and data moats over years, with significant investment of time and capital. Those are real edges. So, where does AI actually fit? It's a tool to help you behave better. Not to generate alpha. What it can do is help you build and test strategies, then execute them systematically, removing emotion from the equation. That means fewer panic sells, less revenge trading, and more consistency. What it can't do is turn a bad strategy into a good one or create a magic money tree. This is still an edge, just a different kind. AI can make you more disciplined, but not smarter. And if you think about where most trading losses actually come from, that distinction matters more than people realise.

@IceSolst There are 2 sides of the AI coding coin and @garrytan and @karpathy represent each. One is putting in a case for high LOC agentic coding and the other is simplifying LLMs to 200 LOC. Only one is right! Choose wisely!

To all the young folks learning about CS/tech, please keep in mind: Garry Tan is a fucking idiot, listening to anything he says will derail you

@LinusEkenstam Built an ML engine that can predict future health composite scores. Used claude to run the entire autoresearch loop that brought the model to the needed statistical thresholds with anti-overfitting guardrails. 20 prospective customers already. districtmap.netlify.app

Can I see some links of what you’ve built with Claude code? Blow my mind

@Chris_Orlob Love these nuggests. For someone trying to breakthrough into AI native SaaS these are precious than Iranian Oil now!

Gong grew from $200k ARR to $200M ARR and $7.2B valuation in a 5 year span. Buyers told us our demos were 2nd to none. 9 lessons I learned about SaaS demos I'll never forget:

@karpathy Have been sitting and working on this concept taken to the stretch of how the nature of news media should evolve. Debating agents (truth seeking) over a news to deliver a balanced depth. I have seen merit in thr first few prototypes.

- Drafted a blog post - Used an LLM to meticulously improve the argument over 4 hours. - Wow, feeling great, it’s so convincing! - Fun idea let’s ask it to argue the opposite. - LLM demolishes the entire argument and convinces me that the opposite is in fact true. - lol The LLMs may elicit an opinion when asked but are extremely competent in arguing almost any direction. This is actually super useful as a tool for forming your own opinions, just make sure to ask different directions and be careful with the sycophancy.

@akoratana @matei_zaharia @pbailis @rauchg @zoink @drewhouston 100X, and this looks promising! All the best!

Introducing: PlayerZero The world's first Engineering World Model that puts debugging, fixing, and testing your code on autopilot. We've raised $20M from Foundation Capital, @matei_zaharia (Databricks), @pbailis (Workday), @rauchg (Vercel), @zoink (Figma), @drewhouston (Dropbox), and more PlayerZero frees up 30% of your engineering bandwidth by: 1.⁠ ⁠Finding the root cause for bugs & incidents in minutes that engineering teams take days to identify. 2.⁠ ⁠Predicting in minutes, edge case issues that a 300-person QA team would take weeks to find. ------ Here's why this matters: No one in your org has a complete picture of how your production software actually behaves. Support sees tickets. SRE sees infra. Dev sees code. Each team builds their own fragmented view - and none of these systems talk to each other. When something breaks, everyone scrambles to stitch the picture together by hand. PlayerZero connects all of it into a single context graph - → The Slack thread where your lead said "we went with X because Y fell apart in prod last time" → The PR review where an engineer explained the tradeoff → The lifetime history of your CI/CD pipeline, observability stack, incidents, and support tickets So you can trace any problem to its root cause across every silo. And it compounds. Every incident diagnosed teaches the model something new. The longer it runs, the deeper it understands - which code paths are high-risk, which configurations are fragile, which changes tend to break which customer flows. So when you sit down to debug a live issue, you have your entire org's collective reasoning and production memory behind you - instantly. ------ Zuora, Georgia-Pacific, and Nylas have reduced resolution time by 90% and caught 95% of breaking changes and freeing an average of $30M in engineering bandwidth. ------ Our guarantee: If we can't increase your engineering bandwidth by at least 20% within one week, we'll donate $10,000 to an open-source project of your choice. Book a demo - bit.ly/3NlLMeN

@wh0sumit Need

got 3 claude cowork invites. lmk if you want one, DM's open 🫡