Researchers show that Claude Code is 98% not AI. Anthropic never gave us the architecture for Claude Code. There were no docs. Just a tool that every developer is currently obsessing over. Until it leaked recently. A research team pulled the source code, analyzed all 500,000 lines, and found something ridiculous. Only 1.6% of the codebase actually interacts with the AI model. The core of Claude Code is literally just a simple while-loop. It asks the model what to do, runs a tool, and repeats. So what is the other 98.4%? It is hardcore, traditional software engineering. The researchers found a massive, complex infrastructure designed entirely to babysit the AI and keep it from hallucinating or destroying your computer: - A 7-mode permission system acting as a security bouncer. - A 5-layer context compaction pipeline so the AI doesn't forget its goal. - A subagent delegation mechanism with strict worktree isolation. - Four different extensibility hooks to manage external tools safely. Every startup right now is trying to build a better AI model to get better results. Anthropic did the exact opposite. They took an existing model and built a fortress of deterministic software around it. They realized that the AI doesn't need to be smarter. It needs to be managed.

Love this stunned ⁦@ChrisCuomo⁩ take on David Grusch’s candid statement that the US Govt absolutely knows we are not alone. It’s a very fair question, why hasn’t someone from the Govt come out and denied Grusch’s claims if it’s not true?

Talk about an absolute nightmare! 🤣🤣🤣

In 1987, American Airlines saved $40,000 by eliminating one olive from its First Class salads.

There are no new programming trends.

Claude Code Mythos and Fable is INSANE when you use ultracode effort. You can become a millionaire overnight. Here's how 👇 step 1: Start using it as a billionaire

The average Claude Mythos user be like

More AI-generated code doesn't make your team faster. It might actually slow you down.

This brought me entirely too much joy. 👽

God forbid Bob Lazar starts feeling himself.

Disturbingly high incidence of the output from this being teletubbies holding shotguns which really makes you wonder about whats in the training data

You don't have to choose between being an AI doubter or a TokenMaxxer. it's ok to be in the middle. I find it more interesting to look at both sides and make my own opinions based on experience (with this tech and many before it). the religious fervor on both sides is exhausting.

Once the bubble pops, Anthropic and OpenAI will become the Coinbase and Block of the AI world. Mundane companies that ship narrative wrappers on mundane bytes. That the bubble will pop isn’t some apocalyptic doomsday prophecy. It’s not that complicated: AI is freakishly expensive to serve. If the returns on the other end are not justified, the bubble pops. And thus begins the decades long buildout to actually economically justifiable AI. It’s amusing how resistant reality is to our fictions and fantasies. In the peak of the crypto bubble we thought reality was going to be transformed into financial liberty and democratization for all, and network states and decentralized reserve currencies. Coinbase stood to be a multi-trillion dollar company and is now just a mundane tech startup. Today we spin similar narratives about the intellectual upheaval of AI, about the new democratization of intelligence and how everything will soon begin to orbit this new technology. At the end, Anthropic and OpenAI will be mundane IT providers with an insanely grim research outlook to make AI economically sensible and useful, no different from Google’s position in trying to make quantum commercially viable. Reality is, fortunately, pretty hardened against our delusions.

Coding is basically the pinnacle of what you could reasonably automate with AI, and yet we still need human engineers to oversee agents for them to be effective. The AI models are trained on an incredible amount of sophisticated code. The users are highly technical and can use the latest tools quickly. The work is “verifiable” because you can test an app. The outcomes are often removed from the quality of the code (you can have sloppy code but the app can still work). And the context for the agent is often already digitized and sitting in the codebase. That’s an incredible amount of benefits that AI coding agents get to work with. Some of those apply to knowledge work, but most don’t in areas where the work needs to be fully reviewed to be useful, or where data isn’t as abundantly digitized. This makes the job for agents in knowledge work more complicated. So if with all of that, engineers still remain in very high demand, the risks are going to be less than what’s perceived for other areas of knowledge work. Agents will let people do far more than they did before, but the people don’t go away.

🚨🇺🇸 YouTubers staking out Area 51 may have just caught the first wild footage of the F-47, America's sixth-generation fighter. If real, this is the plane built to dominate skies for the next 40 years.

I am so thoroughly convinced that anyone who thinks AI 100x's their output is a liar or a lunatic. You are telling me you can make 1 years worth of decisions in 3.65 days? Let alone describing those accurately and coaxing the result from the AI... (1.8 days european time)?

who could have seen this coming? 🤷‍♂️

BREAKING: CEO discovers tokens cost money