New: An "expert witness" in a $61 million lawsuit over an industrial explosion that killed three people and destroyed 200 homes used ChatGPT to write his report for the court.
He prompted ChatGPT to "show how 3M is 0% at fault for the explosion at Watson Grinding"
Introducing Claude Sonnet 5, our most agentic Sonnet yet.
It makes plans, uses tools like browsers and terminals, and runs autonomously at a level that just a few months ago required larger and more expensive models.
i now create more unfinished projects in a month than in my entire cumulative career.
i can't tell if it's a good thing that i'm able to quickly see my ideas are trash, or if the process itself makes every idea feel like trash.
there has been a gradual vibe shift:
- less talk of AI hallucinations and psychosis (psychosis was a gpt-4 era problem)
- less talk of whether AI will replace or augment humans (augment won)
- less talk of whether AI is intelligent (it is a class of intelligence)
- less talk of AI consciousness (who tf knows/cares)
discourse might just be dictated from the top, since most good content comes from reacting to whatever Anthropic is agitating us with.
but in a sense AI feels less like an alien invader now and more of a mundane occurrence.
Agents can deal with ambiguity better than humans in localized cases. It’s a bit uncanny.
On a global scale however agents dealing with ambiguity will trash your codebase.
You must refactor towards human principles of goodness.
Refactoring is the act of changing the structure of code to be more consumable. But consumable by whom? If the audience is humans, we re-factor to human standards. If the audience is agents, we re-factor to agent standards. Those standards are different.
When agents are the audience, I tolerate slightly larger functions, and slightly more comments. This is because agents are more able to deal with larger functions (cyclomatic complexity less than or equal to six). Agents also use comments well. They actually read them. And they maintain them.
These differences are not large, but they are real.
Elon's AI won.
0:00 - Your digital stripper
1:56 - Grok beat Claude.
4:34 - The Doom Trolling Needs to Stop
8:22 - A face only a mother could love
10:45 - AAA gaming is dying
Tim Ferris just shared that he's on track to sell 80% fewer books in 2026 than he did in 2022
"“how-to” books are getting crushed because LLMs seem to provide faster, cheaper, and more personalized advice"
"I don't think people managers will have any value in the future."
Airbnb CEO said on a recent pod the people most screwed by AI will be people managers.
Perhaps we had this all wrong. The future belongs to ICs. There is now more use for them than ever before.