Yakob Vaccaro

Building Kalcine. challenge yourself to be more. | @columbia cs & econ alum

[Yakob Vaccaro @_v_jakovac]

@chamath Facebook?

[Yakob Vaccaro @_v_jakovac]

Only A24

[A24 @A24]

YOU CAN SEE EVERYTHING, a documentary by Nathan Fielder and Lance Oppenheim. Featuring Elizabeth Holmes. Only in theaters this October.

[Yakob Vaccaro @_v_jakovac]

@signulll This is risky if one’s not grounded, but you should be fine. I remember people getting stuck in an infinite suggestion-confirmation loop with 4o.

[Yakob Vaccaro @_v_jakovac]

They didn’t disappoint

[Yakob Vaccaro @_v_jakovac]

gpt 5.5 in codex is so good that I’m, for the first time in a while, excited to see a model release from @OpenAI

[Yakob Vaccaro @_v_jakovac]

Are we going to keep moving the goalposts or is it time to admit AGI is staring right at us?

[Yakob Vaccaro @_v_jakovac]

@humfhuang What you’re not accounting for is that they get infinite attempts on target. Building is easy, tech is improving every day, and they’ve already got name recognition.

[Yakob Vaccaro @_v_jakovac]

@startupoppa That sounds like hell.

[Yakob Vaccaro @_v_jakovac]

It’s that time of the season when I migrate back to Codex from Claude like a bird.

[Yakob Vaccaro @_v_jakovac]

anddd it was glm 5.3 Flash. - “all of this traffic served on Chinese AI chips”

[Yakob Vaccaro @_v_jakovac]

@willyhopps @A24 @osmo_studio Congrats Will !! Exciting

[Yakob Vaccaro @_v_jakovac]

Ridiculous 😂

[signüll @signulll]

lol

[Leonardo @mrloldev]

2 - 0 btw

[Yakob Vaccaro @_v_jakovac]

The best team has won. All is well. 🇪🇸

[Yakob Vaccaro @_v_jakovac]

Insane. Anthropic’s pricing model makes more sense by the day. They were trying to stack it up before this inevitably happened, I guess 😂

[Kimi.ai @Kimi_Moonshot]

Introducing Kimi K3: Open Frontier Intelligence

🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost
🔹 Built for long-horizon agentic coding and self-evolving workflows

Kimi K3 is now live on on https://t.co/zrk6zZxZUo, Kimi Work, Kimi Code, and the Kimi API.
Open Weights by July 27, 2026.

🔗 API: https://t.co/XCrgjXAqMw
🔗 Tech blog: https://t.co/YTfiMSNM1f

[Yakob Vaccaro @_v_jakovac]

Promising

[Thinking Machines @thinkymachines]

Today, we are introducing Inkling.

Inkling reasons efficiently across text, image, and audio modalities. We are making the full weights available.

https://t.co/Ghebq5mG30

Available today for fine-tuning on Tinker. Play with it in the Inkling Playground. 🧵

[Yakob Vaccaro @_v_jakovac]

This is important to me.

[Harj Taggar @harjtaggar]

@mahbodshahshah1 It’s ok to not use Steve Jobs analogies for everything ❤️

[Yakob Vaccaro @_v_jakovac]

Interesting

[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.

[rahul @0interestrates]

guy selling post trained oss models: if you don’t own your models you don’t own anything

guy selling harness on top of models: if you rely on one model you are cooked

guy selling tokens: instead of manually prompting your models just put them in a loop where they prompt themselves

[roon @tszzl]

you either die a frontier lab or live long enough to see yourself sell compute

[Yakob Vaccaro @_v_jakovac]

gpt 5.5 in codex is so good that I’m, for the first time in a while, excited to see a model release from @OpenAI