Everyone in the open LLM community knew this was coming.
We didn’t know the exact timing, but OpenAI is completely stagnant, and it was coming this year or the next.
I don’t think the world still understands how screwed OpenAI is. It isn’t just that their moat is gone, it’s that, even with all that money, their models (for the size\investment) are objectively bad.
I suspect china is fudging the training timeline tho…
I’m more prone to believe OpenAI is just a clunky POS. DeepSeek released a model that’s operating on theories kicking around the LLM community for years. Now Alibaba is claiming they’ve got a better model, too.
Altman insisting he needed $1T in new physical infrastructure to get to the next iteration of his product should have been a red flag for everyone.
They’re trying to brute force a solution to a problem that more elegate coding accomplishes better.
I had suspicious before, but I knew they were screwed when Qwen 2.5 came out. 32Bs and 72Bs nipping at their heels… O3 was a joke in comparison.
And they probably aren’t fudging anything. Base Deepseek isn’t like crazy or anything, and the way they finetuned it to R1 is public. Researchers are trying to replicate it now.
Everyone in the open LLM community knew this was coming.
We didn’t know the exact timing, but OpenAI is completely stagnant, and it was coming this year or the next.
I don’t think the world still understands how screwed OpenAI is. It isn’t just that their moat is gone, it’s that, even with all that money, their models (for the size\investment) are objectively bad.
Yeah it went from hey the monopoly justifies the cost. To Oh shit they did it for how much? Real fast.
I suspect china is fudging the training timeline tho…
I’m more prone to believe OpenAI is just a clunky POS. DeepSeek released a model that’s operating on theories kicking around the LLM community for years. Now Alibaba is claiming they’ve got a better model, too.
Altman insisting he needed $1T in new physical infrastructure to get to the next iteration of his product should have been a red flag for everyone.
They’re trying to brute force a solution to a problem that more elegate coding accomplishes better.
I had suspicious before, but I knew they were screwed when Qwen 2.5 came out. 32Bs and 72Bs nipping at their heels… O3 was a joke in comparison.
And they probably aren’t fudging anything. Base Deepseek isn’t like crazy or anything, and the way they finetuned it to R1 is public. Researchers are trying to replicate it now.
Also, the thing the Chinese govt did probably do is give Deepseek training data.
For all the memes about the NSA, the US govt isn’t really in that position, as whatever the US govt has pales in comparison to Microsoft or Google.
And even as OpenAI struggles, they set fire to a lot of money in the process: https://www.notebookcheck.net/GPT-5-development-hits-major-setbacks-as-OpenAI-runs-out-of-training-data.937049.0.html