Good read. The big elephant in the room though: you likely won't purely hand-code the Stable version for much longer. So where's that split? Prototype vs. Prod? Feature Flags? Canary? 2 codebase nightmare? All of this already exists.
The message that hits for me is that of AI being a destabilizer while simultaneously being an accelerator. The Speed/Scale suggestion won't address this. A codebase no one understands, growing at machine speed won't go away just because you drew a box around it. The fix is likely more mundane stuff like process and role shifts, smaller PRs, tests, tooling, ownership principles.
What I actually would be more willing to allow is a version of this that is built into macOS, runs locally, never phones home. If Apple Intelligence made an "AI sees everything on my screen", I might turn it on.
The specifics are crazier than most people realize. During the rename chaos, Steinberger called Altman to check if "Open"Claw was safe to use. That one phone call became the relationship that became the deal.
Before the forced rename, Claude was the default model, and Anthropic's own Constitutional AI research had inspired soul.md. The trademark enforcement accidentally created the introduction.
The part not getting enough attention is the governance gap, like the MIT license, foundation format, no IP transfer, etc. Good on paper. But the foundation hasn't formed yet, no board members, no governance documents, no clarity on trademark ownership or on the contractual rights OpenAI holds.
This is a brilliant, Love that timer, not sure why the exact time frame was picked though.
It makes me want to create a similar landing page for Apple to fix Spotlight Search. I remember when I used to be able to just find and launch apps on my Mac.
Author here. The earnings call quote gets all the attention, but the engineering blog tells a different story. Their staff engineers talk about faster migrations, not replacement. The gap between the investor pitch and the eng reality is where the interesting bits are.
The really interesting part is what this system runs on.
Spotify built Fleet Management back in 2022 to apply changes across thousands of repos. Half their PRs were already flowing through it before any AI was involved, but only for mechanical stuff like dep bumps and config updates. Claude Code is what let it understand what the code means, not just its structure.
The thing nobody's talking about is that none of this AI automation works without Backstage doing the boring work of tracking who owns what, how things build, and what depends on what.
It it was a bit surreal to see it happen live. GH project went to 70k stars and got a trademark cease‑and‑desist from Anthropic, had to rebrand in one night and even got pulled into an account takeover by crypto people.
OP here, one thing I'm still curious about is if Anthropic's trademark hit was just bad/good timing or if the viral growth triggered things over the edge Guessing it was the latter. That project had been called Clawdbot for months before it blew up. Would anyone care if it has hovered around 5 K stars?
also, ran a Shodan search while writing this—994 exposed instances, mostly on Hetzner and DigitalOcean. if anyone here was running one or got hit by prompt injection, would love to hear about it.
random idea, instead of blocking scrapers, maybe detect them (via user-agent, request patterns, ignoring robots.txt) and serve them garbage data wrapped in dad jokes.
if (isSuspiciousScraper(req)) {
return res.json({
data: getDadJoke(),
artist: "Rick Astley", // always
album: "Never Gonna Give You Up"
});
}
Feels like this is solving a problem that /compact should solve but doesn't. The fact that post-compaction Claude 'feels dumber' suggests the summarization is too aggressive? Would be interesting if Anthropic exposed more control over what gets preserved vs. compressed ... or let users provide their own summary template.
The real story isn't that Google Ads stopped working, it's that attention moved somewhere ads can't follow. You can't buy placement in a Discord server or an iMessage group chat. If this is the new normal, the entire ad-supported internet is running on borrowed time.
+1 also nobody felt guilty for not enjoying NFTs. With LLMs, there's this weird pressure to not just use them, but to like using them. Opting out of joy is supposed to be the price of staying relevant.