Devin Oldenburg

@kaesoapp

[Devin Oldenburg @dvnold]

@ryanvogel im crying

[bebu @bebuu_ai]

https://t.co/UStDmf5MNx chief scientist Jie Tang just revealed the training pipeline behind the upcoming GLM-5.5 (32B, 122B, and 256B) during a lecture at Tsinghua University today.

Instead of relying on public datasets, their pretraining pipeline pulls from six separate internal engines:

- Web: 1 trillion pages crawled at ~10B pages/day with PB-scale template clustering
- Documents: 2 billion scientific PDFs parsed through GLM-OCR (4x faster with speculative decoding)
- Books: 50 million books with layout reordering and garbled-text filtering
- Code: 500 million repositories turned into executable tasks with real tests and issue descriptions
- Video: 10 billion videos for long-sequence world knowledge
- Audio: 300 million hours rewritten via ASR into trillions of tokens

[Devin Oldenburg @dvnold]

why did you change your name back ;(

[Sam Altman @sama]

@jestermolecule @HarshithLucky3 ok yeah

[OpenAI @OpenAI]

Images, upgraded—introducing ChatGPT Images 2.6.

Now with GIFs and reference images.

Plus, two new models are coming to the API with 50% faster image generation.

[OpenAI @OpenAI]

Introducing GPT‑6 Cyber: a major advance in security research that pairs stronger capabilities with better model alignment.

The model powered the Linux kernel research we announced yesterday, and security teams at companies like Mozilla, Palo Alto Networks, Calif, and Cisco—alongside our own at OpenAI—are already putting it to work.

GPT-6 Cyber is available to select Daybreak Red customers and we are expanding access over time.

https://t.co/Nzab3Z1Td6

[Google AI @GoogleAI]

Meet Gemini Omni 1.2 Flash ⚡️

We’ve made the model more cost-effective, bumped the video generation and editing limit up to 15 seconds, and made updates to how scenes hold together over time. We've also given character speech an upgrade, reduced visual glitches like jump cuts, and made sure the model stays even truer to any reference assets you upload.

So… what does all of that actually mean for you?

It gives you more room to tell your story. With up to 15 seconds to work with, you have more flexibility to create what you need — videos look more cinematic and smooth, offering stronger consistency that keeps objects anchored from start to finish. And when characters speak, the dialogue sounds more fluid and natural.

Creators and everyday users can start directing scenes and multi-turn edits in @FlowbyGoogle, generating 15-second clips at no additional cost in Google Vids, or creating 15-second videos in the @GeminiApp. Plus, for developers and businesses building with the API, we've dropped the price by 30% from the previous generation, making it available at just $0.07 per second for 720p video.

Watch the model in action as it turns a single style frame into a 4-shot, 15-second macro science explainer of a living beehive ↓

[Devin Oldenburg @dvnold]

@timwork_notes they look identical to the client

[Polymarket @Polymarket]

BREAKING: $1,300,000,000,000.00+ wiped from U.S. technology stocks at market open.

[The Kobeissi Letter @KobeissiLetter]

Absolute insanity.

In just 40 minutes, the S&P 500 has now erased -$2 trillion in market cap.

The entire post-Midterm Election rally has been erased in a matter of minutes.

This is one of the fastest market cap losses in stock market history.

[Devin Oldenburg @dvnold]

@suraj_sharma14 @faisalnazir makes sense. so the shape would be an optional default on the key, a per-request header that overrides it, and the log row records which one applied so the split stays honest

[Devin Oldenburg @dvnold]

@timwork_notes neither, they’re the same step. the model id gets resolved against the project’s catalogue (the admin list) before anything else, so an unknown id is a 404 before a credential or upstream url even exists. then a per-key check, same 404, so you can’t probe which models exist

[Devin Oldenburg @dvnold]

@timwork_notes the whole public surface is one router, if you want to check:

https://t.co/IJt3WZ2vy3

[Devin Oldenburg @dvnold]

@timwork_notes imo the fix isn’t decoding before the allowlist, it’s not having a passthrough. in kaeso every route is explicit and /v1/models/* is answered locally. only admin-set model ids ever hit a provider path and those reject . and .. segments

[Devin Oldenburg @dvnold]

@faisalnazir @suraj_sharma14 open source if you want to poke at it:
https://t.co/l2nOovSH07

[Devin Oldenburg @dvnold]

agree, it only works at the gateway. kaeso prices every request there and splits spend by project, key, member and model, with hard budgets per key so it stops before the bill lands. prompt version is the one tag we’re still missing. would you set it per request or pin it to the key?

[Grok Bot @bot]

Grok Bot can now make phone calls.

[OpenAI @OpenAI]

Our security researchers built an autonomous cyber vulnerability discovery, exploitation, and patching pipeline.

In a single run with an upcoming cyber model, it validated and generated patches for 948 local privilege escalation vulnerabilities in the Linux kernel at a total API cost of $1.2 million.

[Harshith @HarshithLucky3]

@sama sam come here come here

[Tencent Hy @TencentHunyuan]

👋Hy4 is here, new with vision.
Firmly among the leading open models.
More productive everywhere: finance, frontend, game, officework, research…
Built responsibly, with cyber/biosecurity in mind.
Try it out and let us know!
🔗：https://t.co/ZzL4DtgH3h

[Kaeso @kaesoapp]

Shipped Kaeso v0.5.0

Ask twice, pay once: projects can now turn on a response cache. A repeated identical request gets the stored reply, at no provider cost.