Jeffrey Emanuel
@doodlestein
Former Quant Investor, now building @lumeraprotocol (formerly called Pastel Network) | My Open Source Projects: https://github.com/Dicklesworthstone
You can read it here— let me know what you think: youtubetranscriptoptimizer.com/blog/05_the_sh…
So all these nice Python libraries like browser-use (which basically wraps Playright browser automation with an agentic AI layer for automated control) have already been nuked by Cloudflare anti-AI scraping measures before they even got a lot of traction. But it turns out there’s…
So does it still make sense for Anthropic to lose money on heavy users like this? Will the additional usage data and understanding they get about real-world patterns of usage for agentic coding flows give them some kind of insurmountable competitive moat versus other labs? I…
So crazy for Google to burn goodwill like this when they are printing money the way they are. They really can’t spare the funds for this? Or the Internet Archive can’t chip in somehow? This is like Geocities 2.0
Wow, these guys just don’t care. This is like grade F internet stewardship. Even if it is expensive for them to keep it going, they can afford it. It just reinforces the lesson that you should never ever rely on them for anything if you care about longevity. Anyway, Bitly FTW.
Is it weird that I trust your pelican test more than all the official fancy benchmarks? It’s the one test you can be reasonably confident they didn’t game/manipulate and train on test.
This same cascading sense of realization and grief is coming like a tidal wave for nearly all knowledge based professions. In a sense, he’s fortunate because it’s hitting math and coding first, so he has a big head start on adapting and reconciling himself to it before most do.
the openai IMO news hit me pretty heavy this weekend i'm still in the acute phase of the impact, i think i consider myself a professional mathematician (a characterization some actual professional mathematicians might take issue with, but my party my rules) and i don't think i…
Someone should really do this as a startup. “Agent Friendly” versions of all the most common utilities like awk, sed, find, ripgrep, jq, that are fully native and nearly as fast but which won’t allow a rogue agent to screw things up too badly (or a malicious agent to exfiltrate).
I tried writing a “safe wrapper” around tools like ripgrep and sed/awk. It’s an incredibly challenging thing to do. I think there’s actually a market for “neutered” versions of popular utilities specifically for use by AI agents.
Claude made me lol today. I gave it a bunch of stuff to fix last night before I passed out from a long coding session, and today I woke up and checked it, and it fixed it all. Plus it assured me that the code was no longer a "BROKEN PILE OF SH*T."
Andrew Huberman is the “Teflon Don” of disgraced podcasters. He should start his own PR consultancy on the side because he clearly knew something others didn’t!
So extraordinary how he essentially ignored that article completely, never responded to it, never apologized publicly or prostrated himself for the ones calling for his cancellation, and managed to come out the other side relatively unscathed except for the jokes. But no marriage