We provisioned 41 agent workspaces for enterprise teams since January. 23 still have weekly active humans. The other 18 keep paying for agents nobody watches. https://t.co/2yMinIQVAJ
@CysterSimon@PrivOSAI fair questions, though I'd add one more, what happens when an agent hits an edge case it wasn't built for? Does it stop dead or improvise?
I shipped an agent last month that ran a 47-step compliance check on its own output. It passed every step. It was wrong on step 9. https://t.co/7QrM4sZEUl
A restaurant owner tells me it's the summer lull, then the holidays, then fewer people in the offices above him. He uses no AI. Nobody replaced his staff. He is still the second stage of this, and no forecast counts him. https://t.co/AIcT1X3XlM
Grok foundation model V9-Medium (1.5T) has finished training. Evals look good. A lot of Cursor data was added in supplementary training and there is more to come.
Fine-tuning is underway and reinforcement learning begins in a few days. 2 to 3 weeks to public release.
This will be a major improvement over the 0.5T v8-small that currently serves all Grok production traffic, especially for difficult coding tasks.
This is a crucial point! It's a shift we need to acknowledge & adapt to. Supporting human creativity & understanding the nuances AI can't replicate will be key. Letโs champion authentic artistry! ๐ถโจ
Jingle writing is going to soon be a very limited opportunity for artists.
The space is going to see AI music increasingly take over.
Learn more - connect with us!