JUST IN: OpenAI officially releases GPT-5.6 limited preview, claiming to be more powerful than Claude Mythos.Key improvements include a much larger 1.5 million token context window, better agentic capabilities, faster performance, and fixes to previous issues.Due to its advanced power, the rollout is limited to government-approved enterprise customers for now. A wider release is expected in the coming weeks.

Sam Altman says the current AI pipeline works, but the breakthrough will come when AI can help solve that research problem itselfOpenAI aims to deploy 500,000 A100 GPUs as an AI research intern by September"then build an end-to-end researcher capable of discovering new architectures by march 2028"

OPENAI JUST OFFERED THE US GOVERNMENT A 5% STAKE.🇺🇸 OpenAI has proposed giving the US government a 5% stake in the company, worth $42.5 billion.According to the Financial Times, the stake would help fund a new public wealth fund, allowing Americans to benefit from the growth of AI.The proposal is still in the early stages, no agreement has been reached.

OpenAI Noam Brown:I’d be surprised if our latest internal models don’t get a perfect IMO score — and disappointed if we don’t release a public model this year that can do the same"either now or very soon, math and coding competitions won't be interesting anymore"The frontier will be unsolved problems and real-world research

OpenAI President Greg Brockman:"every time we've run into 'the model isn't scaling the way we expect, ' it turns out we have a bug — our math wasn't right, or the implementation didn't match the mathThat's an important thing to internalize

This is the MOST DEVASTATING development in the fight against AI totalitarians ever.It makes "the Deepseek moment" look like child's play.In a world where closed source labs like @OpenAI, and @AnthropicAI collude with the USG to create a cartel of central planning for cutting edge models, this brilliant end around promises to upend everything that the AI monopolists (looking at you @DarioAmodei) are working towards.The problem with 1 super powerful frontier MoE model is that it has a single thro