Jesse Heikkilä retweetledi
it’s surprising to me how many people seem to not understand that great models are built with super high quality curated data
finding novel ways to create / get this data is a huge edge
English
Jesse Heikkilä
338 posts
Jesse Heikkilä
@vcjesse
Founder & VC at Failup Ventures. Vibe investing in early-stage companies in the US and Nordics. Previously founder/cto/product
San Francisco Katılım Temmuz 2015
467 Takip Edilen116 Takipçiler
Jesse Heikkilä retweetledi
AfterQuery is quickly closing the gap on the
“giants.”
We’ve already grown multiples past the $100M revenue run rate cited here.
If you’re a post-training researcher or lab that needs long-horizon tasks, professional / knowledge work data, high-fidelity RL environments, code generation data, or internal evals done right, my DMs are open or reach out to us at research [at] afterquery [dot] com.
Deedy@deedydas
Every single startup selling AI Training Data (July 2026) >50 cos sell data and RL environments to big AI labs and drive AI progress behind the scenes. They total ~$8.5B in rev and ~$100B in valuation, >75% of which are just 4 players: Scale, Surge, Mercor and Handshake.
English
Buying time to release a new model? Regardless, OpenAI 5.6 release shows how competition is very important for customers of products.
Claude@claudeai
We're extending Claude Fable 5 access on all paid plans, as well as keeping Claude Code’s weekly rate limits 50% higher, through July 19.
English
Interesting release. If this still was open source it would be clear SOTA most likely and get very fast adaption.
Mark Zuckerberg@finkd
(1) Today we're releasing Muse Spark 1.1 -- a strong agentic and coding model at a very low price. It's available through our new Meta Model API and in Meta AI.
English
Jesse Heikkilä retweetledi
@mehulmpt so much cheaper than the estimates Jarred had and a few OOMs faster.
English
It took $165,000+ of Fable 5 API to port Bun from Zig to Rust.
Wow!
Jarred Sumner@jarredsumner
Rewriting Bun in Rust bun.com/blog/bun-in-ru…
English
I think @RoKhanna has some good points with the centimillionaire tax and Bucco is onto something.
No company should be valued at above $10B. Pure greed. Companies above that market cap threshold should have to pay 1% revenue tax per day until they are compliant.
BuccoCapital Bloke@buccocapital
Just make it illegal to be a private company once you reach a $20B valuation. Have to IPO. Would basically fix 99% of the shit people complain about. Enough already
English
Jesse Heikkilä retweetledi
Just make it illegal to be a private company once you reach a $20B valuation. Have to IPO. Would basically fix 99% of the shit people complain about. Enough already
English
@bridgemindai Everything? You mean until the next SOTA coding model comes out from whoever?
English
GPT 5.6 launches next week.
The SAME WEEK Anthropic pulls Fable 5 from subscriptions.
And for the first time, OpenAI actually has a shot:
- GPT 5.6 Sol Ultra already beats Fable 5 on TerminalBench: 91.9% vs 84.3%
- Half the price
- Less guardrails killing normal coding tasks
- No weekly limit that serious builders burn in 24 hours
Fable 5 is the best model on earth, but the guardrails and rate limits are nerfing it.
GPT 5.6 is going to have less guardrails and the rate limits will be WAY better.
Next week decides everything.
English
This definitely goes to goals I hope to do with my son. Just have to wait a few more years.
Justin Uberti@juberti
Pair programming back in the day
English
Jesse Heikkilä retweetledi
Great to see more Finnish tech companies list on the NASDAQ. Congratulations to IQM!
TechCrunch@TechCrunch
IQM, Europe’s first public quantum company, admits the future of the tech is uncertain techcrunch.com/2026/07/02/iqm…
English
Jesse Heikkilä retweetledi
Levels of Individual Agency:
L0 — I don’t know how to do this.
L1 — IDK. Please teach me.
L2 — How do I teach myself?
L3 — I did the task. Can you check it?
L4 — I did the task well. What’s next?
L5 — I did all the tasks.
L6 — …silence…
English
This is all the sun I need. Release it, I'll pull the curtains down!
OpenAI@OpenAI
GPT‑5.6 Sol sets a new state of the art on Terminal‑Bench 2.1, which tests complex command-line workflows requiring planning, iteration, and tool coordination.
English
@thsottiaux Ask User tool to be usable outside of Plan mode! Efficient way of looping over decisions/scenarios mid-task as well.
English
Why is this being reported as a bad signal? If this was a startup who spent $3.7M out of $77M cash reserves that would be extremely low burn.
Techmeme@Techmeme
Docs: OpenAI burned through $3.7B in Q1, on revenue of $5.7B, and ended the quarter with $73B+ in cash and marketable securities vs. $40B at the end of December (@erinkwoo / The Information) (Visit Techmeme dot com for the link and full context!)
English
Need to dig up all the old Nokia phones for my new data center. Guaranteed 100% uptime
Google Research@GoogleResearch
Today on the blog, we discuss a pathway for the second life of phones through the exploration of “phone cluster computing”, which can directly reduce the environmental footprint of computing by avoiding the need for further raw material extraction. More →goo.gle/4aJe5vO
English