Software Dev • American living in the EU 🇺🇸🇪🇺🇬🇷 • Photographer • I’m fascinated by a future where AI will make being human better

Internet
Now that Opus is back, I’ve settled into using it as my primary code generator, with Sol being my primary reviewer. All Astra all the time was fun for a bit, but this hits my sweet spot of speed, quality, and tone.
4
4
24
94,660
While test driving the UI of the app we’re building, @claudeai keeps using Ada Lovelace as the placeholder person name. 😍
2
1
437
Opus is back, baby. Way too early to say too much more, but it certainly has its mojo back.
2
366
Sol 6 and Opus 5.5 are the pacing I’m here for.
1
1
5
774
This isn’t being snarky, btw. Improvements in the Sol and Opus tier make material changes in our day-to-day.
117
🤣
Since Jev launched Tuesday, there’s apparently a new thing at @every “Jev boys.” Ever since the launch there’s been this flock of self-proclaimed “Jev boys” who carry Jev on hand at all times and constantly ask it what to do. They have their entire personality revolve around Jev, TypeSafe, and calibrated probabilities. When we went around talking about what people were trying this week, 4 or 5 of them said some variation of “I live by the Jev and die by the Jev.” Just about an hour ago, I asked someone to do something and one of the Jev boys was bold enough to say “if Jev says I do it, otherwise I don’t.” I told him if he asked Jev instead of just doing the thing I was going to lose my mind. He asked it anyway and it said to do it. Thank god for that at least. But then the other Jev boy asked the same question and Jev gave him a different answer. He immediately pulled up the probability and started explaining that Jev was only 63% confident, so technically it was a judgment call. I told him that yes. That is what being a person is. Later, another Jev boy asked Jev whether he should go to a meeting. Jev said no. He did not go to the meeting. When someone asked where he was, he posted the Jev output in Slack. At lunch, I watched two Jev boys ask Jev where they should eat. They got different answers and spent five minutes comparing confidence scores instead of picking somewhere. One of them actually said “I’m going with Jev on this one.” The other said “your Jev is wrong.” I don’t even know what that means. Now they’ve started asking Jev questions about each other. Someone asked Jev if another person’s idea was good. false The person whose idea it was then asked Jev whether that was constructive feedback. false Neither of them has spoken since. This afternoon one of the Jev boys told me he doesn’t make decisions anymore, he “routes them through Jev.” I asked if he was joking. He asked Jev. It said false. It has been three days since launch. I really hope this dies off soon. I haven’t seen anything outside Every about “Jev boys,” so I’m guessing it’s just an Every thing. Unless one of you has more information.
1
578
Oh thank goodness.
We're adding support for AGENTS.md to Claude Code. Starting today in version 2.1.277, if there is no CLAUDE.md in a folder, Claude will check for and use AGENTS.md. You can toggle this behavior in /config.
5
429
I know I’m a day late, but I think the mental model I’ve latched on to explain Jev it is that it’s to regular classifiers what a software defined radio is to an old school single-frequency crystal radio. A "software defined classifier" if you will.
2
6
393
There are so many ways to AI. Glad to see this take and looking forward to trying it out.
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI? I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev • 20-200x faster • 40-400x cheaper (w/ output tokens free) • Frontier composable intelligence optimized for decisions AFAICT the shortest path to AI-based economic revolution
4
470
I know, I know. We’ve got to get the cabin ready for landing, but I’m waiting on Fable to be done thinking!
1
4
326
Happy OS 27 day to all those who celebrate.
291
Duncan Davidson retweeted
Sure would be nice if people were at least as concerned about climate change as they are about AI
8
9
53
7,293
If we’re going to worry about p(doom) (and we should), I think we really figure out what Δp(doom) is. After all, we’re already operating under a pretty decent p(doom) from our current non-AI situation (climate change, war, etc...)
2
6
424
I’m think p(doom) from climate change is likely great than that from AI, BTW... and then there’s the political situation and... so much more. None of these is likely to wipe humanity off the face, but all of these can be the end of the world as we know it.
1
134
I spent several hours on a blog post about AI and the fear of human civilization ending that I ended up abandoning. Bah.
1
4
312
Gotta admit, I’m pretty happy to see this. I have been expecting this trend to pick up now that priors for using RN are changing.
Native is now the future of mobile at @Shopify : shopify.engineering/back-to-… We don’t hold on to a decision just because it was successful at the time. When a core assumption changes, we’re willing to go back and ask whether it’s still the right call. LLMs changed one of the core assumptions behind our 2020 decision, so we reevaluated our mobile stack from first principles. What we found led us back to native. The @shop app has already been migrated. How we did it: shopify.engineering/shop-app…
1
2
14
1,897
While critiquing something I’m writing, Sol randomly said: “I’d use... verantwoord?” When pressed, Sol said: “‘Verantwoord’ is semantically adjacent to what I was doing — judging whether the paragraph was defensible or justified — so a Dutch token from that conceptual neighborhood surfaced instead of the intended English transition. It’s a tiny example of multilingual representation leaking through generation: the concept arrived correctly, but the language boundary briefly didn’t. Not a hidden editorial notation—just an unusually legible glitch.” Its suggestion was a good one, by the way. Just an interesting slip.
271
I’ve got a feature request for you @thsottiaux: Can we see model/effort being used by a subagent in the Subagents list or panel? Currently, I have a line in my AGENTS.md to give output so that I know how well my subagent rules work.
280