A2P2

@Mnilax I have not been able to steer its behaviour with any level of restructuring / cleaning up both the files. So I gave up on trying to make it work.

@Mnilax For example - one of the rules that I add is research API docs / latest data before acting (e.g. for infra stuff). But it always falls back to using data from training. This has been particularly frustrating

Karpathy threw a grenade at every senior engineer who still treats LLMs as a toy. his actual words: the worst thing an expert can do right now is reject them. most experts read it as a threat, but it's advice. his framing: > the gap between "AI tools are bad" and "AI tools are useful when used right" is professional discipline, not capability > agents have cognitive deficits. they fail in ways nothing in the training set anticipated > the experts who reject LLMs lose to experts who learn to wrangle them > "models have so many cognitive deficits. but you can route around them" routing around the deficits is what CLAUDE.md was invented for. Karpathy himself wrote 4 rules. across 30 codebases they took my Claude error rate from 41% down to 11%. solid drop. but his rules pre-date the slop era going public. I bolted on 8 more, tuned to the failure modes that surfaced after January. got it down to 3%. a CLAUDE.md does not raise Claude's IQ. it lowers his slop floor. that is the entire game. open the article underneath. the model is not the bottleneck. your config is.

@Mnilax … it ignores. This is based on latest usage in Mar / April with Claude Max plan.

@Mnilax So if you have a defined project and the shape of your tasks is pretty narrow and well-defined (e.g adding incremental features or refactoring) - it works. But if you are doing anything open-ended (e.g. setting up a homelab, conducting exploratory research etc) …. (1/2)

@leeschmidt123 Isn’t consulting exactly that? Throwing around fancy words which actually mean nothing? They fooled the world long before Sam / Dario dis

Met a guy who’s a partner at one of the largest consulting firms in the US and globally. He started talking about AI coding. What followed was a stream of nonsensical AI and coding related buzz words. Recommended “forking claude” (whatever that means), and “use github workflows for everything so the agents can talk to each other” (also whatever that means). And it only went downhill from there. Large businesses pay these people millions of dollars. So much waste and cluelessness out there it’s insane.

@ThePrimeagen i hate this family (play succession ost)

Dear interns, I am sorry for what I said about Terminal App on Mac. It is better than cmd on Windows I am sorry for lashing out I'll be taking the rest of this weekend off Twitter to understand why I behaved this way. Say we are a family or you are fired, ThePrimeagen P.S.: we are doing layoffs because of AI

@sahill_og You were born yesterday kid. Stop yapping. You don't even know what an engineer means.

If Claude Code, agentic workflows, and MCP still mean nothing to you in 2026 This is probably your last year as a software engineer.

@SahilBloom Sure the message is correct - but the intention is not. Not everything should be done just because it's acceptable or not acceptable by people.

Major life hack: Don't complain, ever. Nobody likes a complainer. They drain the energy of everyone around them. It's exhausting spending time around someone who constantly complains about things outside their control. If it’s within your control, go do something about it. If it’s not, you’re just wasting energy thinking about it. Complaining gives too much power to the thing. Take back that power.

@haider1 Right. Now can we add cost, applicability in non clean-room conditions (e.g. benchmarks), adaptability as well? We can’t

bad news LLMs are starting to look like they're hitting a wall

@loycense @AA_isnowonline @Jonathan_Blow The sad part about the internet is any monkey can type. Like this guy above.

@AA_isnowonline @Jonathan_Blow Nothing goes over my head, I am too fast I would catch it

It's been 3 months since the 100x vibers started 100x vibin'! So, post your 25-years-of-work-equivalent project here, so we can signal boost and everyone can celebrate the Life's Work that you did in 3 months. Looking forward to it, Let's Go!!!

@thesilliconjax @Polymarket Sure mate. I hope you never wake up from your delusion. Maybe when price of tokens goes to 10x - 100z you’ll vegetate entirely cause there will be nothing left to burn

@thesilliconjax @Polymarket Being able to generate your cute PoC by typing english is not a proxy for and says nothing about your ability of maintaining a product. When your app is crashing in prod at 3am and your customers are furious - no amount of “please fix this claude” can save you

JUST IN: Cloudflare lays off 1,100+ employees through email as it restructures for the “agentic AI era.”

@Jonathan_Blow Desperate people rationalizing their AI spending and incompetence to run their companies properly.

Even Andreessen is admitting this now...... x.com/pmarca/status/…

@thesilliconjax Sure technically everyone is a billionaire in their own heads.

@witless_duck My 'garbage' system is generating more MRR in a week than your 'braincells' will in a decade. 💸

I just offboarded my entire Polish dev team. The result? A 1,000% increase in velocity and 0ms communication latency. Founders, stop being philanthropists for legacy outsourcing. The "Polish Hack" is a 2010s delusion. Here is why your Warsaw team is your biggest bottleneck. 🧵

@Av1dlive So convenient to construct a senseless overhyped narratives that will sell your product

Jensen Huang (CEO of Nvidia) : “Every engineer is going to have and manage 100s of agents” The future of work is managing and building ai agents This builder gave out the entire roadmap to become a 100x ai engineer in 2026 Bookmark this for the weekend

@Bencera I want to try the thing that you are on

AGI is here. Capitalism is ending soon. The bar for shipping velocity is inhuman. Everyone is racing to build final companies. May the best win.

@JosephJacks_ This is what happens when you forego thinking and instead ignore years of evolution from fishes to ape to human to back to fish again by delegating almost entirely to a cstochastic idiot. Otherwise going “full retard”. Never go full retard.

PREDICTION: Anthropic will surpass Alphabet in revenue by mid-2028. This is not a bull case or an acceleration scenario — it is a continuation of the curve already in evidence. Anthropic’s ARR went from $1B (Jan 2025) to $9B (Dec 2025) to $30B (Apr 2026) — a 3.3x step in a single four-month window, and the curve has been steepening, not flattening. My projection actually assumes deceleration from here: $100B by end of 2026, $340B in 2027, $850B in 2028, $1.4T in 2029, $2T by 2030. Crossover with Alphabet happens at ~$575B in mid-2028, not because Anthropic accelerates beyond today’s pace, but because Alphabet — locked at ~15% YoY in a mature ads-and-cloud business — cannot match enterprise AI’s adoption physics. As @rodriscoll intelligently observed recently, Gemini tokens served grew by only 60% in the last quarter … while Anthropic grew by 10X. Three drivers make the continuation structural, not speculative: customers spending >$1M/year with Anthropic doubled from 500 to 1,000 in under two months post-Series G (these are multi-year expanding contracts with near-zero churn — switching a deployed agent stack mid-flight is operationally untenable); Claude Code is the wedge, not the product, dragging the rest of the platform — agents, MCP, healthcare, biotech — into every Fortune 2000 deployment as an attach point; and compute supply is finally non-binding with the 3.5GW Google + Broadcom deal (2027+), this weeks SpaceX partnership, and 1GW of standing Google capacity for 2026. For most of 2024–2025 the bottleneck was supply, not demand. That constraint is releasing exactly when the demand curve is steepest. The standard objection — “no company has ever sustained this at scale” — applies a software-era frame to a labor-era business. AWS, Azure, and Meta decelerated at $50–100B because they sold tools to the economy. Anthropic is selling cognitive capacity into the economy. The TAM isn’t enterprise software ($800B). It’s labor ($50T+). When the denominator is two orders of magnitude larger, “deceleration at $100B ARR” stops being a law and starts being an assumption. The crossover isn’t a maybe. It’s a function of timing. Mid-2028 is when I think Anthropic surpasses Google.

@WatcherGuru Karma

JUST IN: Coinbase crypto trading remains down for over 5 hours as outage continues.

@NanouuSymeon clear;reset

Prove you're a Developer, with just one word 👇