Jaime O

[Jaime O @aiburstiness]

M3 Ultra being tested 🔥🔥@Apple @exolabs

[Jaime O @aiburstiness]

I genuinely believe OpenAI made a huge mistake with its Pro plans. I’ve been a Pro subscriber since day one, but this feels like a betrayal of loyal users.

According to their latest email, October 29th appears to be my last day with the current 20x usage allowance. After that, I won’t be renewing my $200 subscription.

OpenAI may be winning the race to lower the cost of intelligence through its API, but it’s losing when it comes to subscription value.

And that “gift” of 62,500 credits for Work and Codex? I struggle to see the real value. Why not let us use them as actual API credits instead?

My advice to everyone paying $200 for ChatGPT Pro: give Claude’s $100 plan a try. It’s an incredible deal for half the price.

Loyal customers deserve better, not more restrictions for the same $200. @thsottiaux @OpenAIDevs @claudeai

[Jaime O @aiburstiness]

@thsottiaux Thanks for the Pro credit grant! One clarification: can the 62,500 promotional credits be used to purchase paid weekly resets, or are they only for token-based usage? I previously saw an $80 reset, while $80 also buys 2,000 credits.

[Jaime O @aiburstiness]

@bridgemindai People should benchmark these metrics: reasoning juice per task and usage burn rate.
Looking only at the headline limits doesn’t tell the full story.

[Jaime O @aiburstiness]

🚨 Confirmed: it’s a banked reset, and it’s already showing as available on my account 🔥

Huge thanks to @thsottiaux for the reset 🙌 and @codex_resets for keeping us updated 👀

[Jaime O @aiburstiness]

@eli26902 @jun_song This will happen just like with London housing: the government approves building projects, but greedy developers build and sell slowwwwly to keep prices high.

[Jaime O @aiburstiness]

@William5Ht Let’s connect, London 🇬🇧

[Jaime O @aiburstiness]

@OpenAIDevs ASTRA still waiting!!!!!

[Jaime O @aiburstiness]

@jun_song I tried Grok 4.6 for a few hours, and it looks solid. The speed-to-intelligence ratio is amazing! Let’s see what happens with 4.7 👀

[Jaime O @aiburstiness]

@kimmonismus Same here. I can’t believe how bad it is with Computer Use. I use Codex a lot to do things for me while I’m away, and it actually gets the job done. Claude is just bad

[Jaime O @aiburstiness]

@thsottiaux Just wait until I see /cerebras. Don’t you dare turn it on 😬😬

[Jaime O @aiburstiness]

@juminoz @repheon @ForwardEditor Do you turn it off ? Wow

[Jaime O @aiburstiness]

@FishAudio I’ve been working on an STT–LLM–TTS pipeline with MLX, but wow, this looks amazing. Well done!

[Jaime O @aiburstiness]

@theo I complained about it, "Expires 8/11" so use it before that day otherwise bye bye !

[Cline @cline]

DeepSeek silently updated their changelog with a new V4-Flash upgrade 1 hour ago.

Their new Terminal-Bench score is 82.7, a massive +25.8 point leap from its initial April preview score of 56.9.

Currently only available via their API, open weights release will follow shortly.

[Kimi.ai @Kimi_Moonshot]

Releasing the model weights and technical report of Kimi K3.

Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window.

New model architecture: 2.5x the intelligence per unit of compute, not just more params.

Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale.

Model weights: https://t.co/7m7eEg6Y0B
Tech report: https://t.co/yeu6cjpMCT
Tech blog: https://t.co/YTfiMSNM1f

[MiniMax (official) @MiniMax_AI]

Open weights.
Open research.
Open innovation.🫶

Marching for an open future.🤍

[Jaime O @aiburstiness]

@thsottiaux @PaulSolt Omg I knew ittttttt ! 🔥🔥🔥thanks

[Jaime O @aiburstiness]

@boxmining @thsottiaux You have usage management issues. 😅 I have more usage left than you!

[Rohan Paul @rohanpaul_ai]

Jensen Huang on "distillation"

On his new interview with axios, he was asked this question

"Should open source model companies be allowed to distill closed models"

"Distillation—learning from AI, learning from other people, and learning from other sources of knowledge, is fundamental to intelligence.

We are constantly learning from other people. I am learning from you through the questions you are asking, and you are learning from me. All day long, we are learning from one another. AI also has to learn from something.

The original AI models, whether they were open or closed, were trained on previously created knowledge from the internet. Now, AI is generating more content than humans. In a few more years, the internet could be 99% AI-generated content, and that content will have been created by some form of AI.

As a result, AI systems will constantly be distilling knowledge and intelligence from other AI systems. The fact that AI can learn is a good thing. We want AI systems to be intelligent because a smarter AI can also be a safer AI."

----

From "Axios" YouTube channel, (full video link in comment)

[Jensen Huang @JensenHuang]

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.

AI will transform every industry, power every company, and be built by every country.

Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.

The world needs both frontier closed models and frontier open models.

https://t.co/AUKzoQ5Ikb