have your agent call my agent. ai-native finance at @Digits. ex-founder @Basis (acquired), @Uber.

Joined December 2011
Modern take on dating : who's taking me to a deprecated models funeral? I'm being serious.
JD Vance: "There are things at some of the AI companies you see happening, where you feel like there is some really, really weird spiritual dark energy. Like one of the companies that develops these models, when they get rid of one model and announce a new model, they actually hold a burial ceremony for the old model, and the employees apparently will cry."
Readers added context they thought people might want to know
JD Vance is referring to one event in which fans held a theatrical “funeral” for the retired Claude 3 Sonnet model in an SF warehouse in early August 2025. The event was organized by fans not any AI company. Only 200 people attended. wired.com/story/claude-3…
1
58
leon retweeted
Replying to @mark_k @AnthropicAI
if true, what a massive waste. we’d have an 80% reduction in token use just by removing the ask_user_question tool
4
1
32
4,904
GLM 5.2 (high) scores 67.3% on WeirdML, a score between Opus 4.5 and Gemini 3 Pro. This is a much higher score than I expected, and GLM 5.2 max (still running) could score even better. It looks like a very solid model!
WeirdML v2 is now out! The update includes a bunch of new tasks (now 19 tasks total, up from 6), and results from all the latest models. We now also track api costs and other metadata which give more insight into the different models. The new results are shown in these two figures. The first one shows an overview of the overall results as well as the results on individual tasks, in addition to various metadata. The second figure shows cost vs performance and shows a clear scaling with better results for higher costs. We also have a very varied pareto frontier with 11 models from 6 different companies having the best accuracy for a given cost for at least some of the cost range. Grok 3, Claude Opus 4 and GPT 4.5 are the ones that underperform for their costs, while Gemini pro and o3 pro have the best results at the highest costs. Qwen3 30B3A, grok 3 mini and deepseek R1 also each represent a good chunk of the pareto frontier.
19
31
14
450
217,064
wow sick new billboards from openai
96
255
96
5,123
2,852,690
leon retweeted
Lots of people asked how I used Fable to edit its own launch video so I made a video about that! TLDR it wrote a lot of code & tool calls to use transcription services, ffmpeg, do colorgrading, use the figma mcp, make remotion UI and render it. I didn't touch a video editor.
293
623
172
9,432
1,101,540
let's gooooo, first mobile app that i built got approved by Apple after 1 rejection here's some random things i learned for other aspiring app builders: - AI is pretty good at researching Apple's guidelines and adhering to them, as well as filling out App Store Connect (use ASC CLI by @rudrank), so it only took 1 rejection - i used Swift which was VERY painful when testing due to lack of hot reload, 100% stick to Expo - for this first app, i made it pretty similar to existing apps so that i could focus on learning about Xcode, App Store Connect, etc, so don't reinvent the wheel at first
7
2
41
1,842
leon retweeted
Replying to @zekramu
b2b saas can be recategorized as private harnesses for open models
1
1
2
46
Two of our worst VC stories: 1. A Sequoia partner passed on Cloudflare because he didn’t think a woman could lead a security infrastructure company. Seriously. 🙄 2. I got introduced to @pmarca. Meeting got scheduled for a Monday, which should have been a clue. I thought it was just a casual meeting. He thought it was a pitch and brought the whole @a16z partnership team. Hilarity ensued. 🤪 At one point one of them said: “You don’t seem very prepared.” Which was true because I wasn’t. I framed the rejection letter they sent.
I was once pitching in a board room at a top 3 VC firm for a $15M Series A. 12 people in the meeting. One of the GPs fully fell asleep. Out cold for 30+ minutes. Nobody acknowledged it. Everyone just kept going. I kept presenting my Series A slides to an unconscious man in a Herman Miller chair and somehow that was considered normal. That's venture capital. You might fly across the country to perform for people who may or may not be conscious. It's a dance. And sometimes you lead and sometimes you follow and sometimes your partner is unconscious. If you're raising right now, just know: every founder has a story like this. The process is weird. The power dynamic is weird. You're not crazy for thinking it's weird. No one talks about it because they want to continue raising. But I'm happy to stick my neck out there. It is weird.
180
259
194
7,449
4,530,484
Tell me GPT 5.6 is coming without telling me GPT 5.6 is coming 👀
Damn, Codex 5.5 has been lobotomized. Today it has the attention span of a gold fish.
28
10
2
482
42,461
leon retweeted
Replying to @claudeai
4.8 finally lets me spin up 100 sub-agents and speedrun my weekly limit
1
167
9,728
leon retweeted
Replying to @TopviewAIhq
how is Seedance 2.0 so frickin good
1
1
35
1,856
leon retweeted
Replying to @evilsocket
dang it Rohan
2
6
471
10,798
leon retweeted
Replying to @thdxr
i posted this elsewhere, they have a caching bug in threads, very obvious in very long-running chats
1
1
43
4,217
leon retweeted
Replying to @OpenAI
sir, do you have a minute to talk about AGI?
1
13
4
309
36,656
leon retweeted
Replying to @petergyang
probably people asking for a rep after getting frustrated with bots
3
1
1
24
6,660
leon retweeted
Replying to @ankkala
Reddit walked so Claude could hill climb
1
2
99
leon retweeted
Replying to @birdabo
oops, ran out of my weekly limit right before we saved humanity
22
37
4,228
124,463
leon retweeted
Codex didn’t cook. It crunched
1
3
419
leon retweeted
Replying to @kimmonismus
my feed is filled with people supposedly switching to openAI, I wonder if that’ll have any impact in next month’s numbers
1
7
496
I’m interested in genetically engineering plants to be much more efficient at photosynthesis and grow faster as a potential way to fight climate change. Anyone working on this?
198
163
46
1,081