MedGemma is a really interesting model – very small, multimodal, open, and does quite well in out-of-distribution medical tasks compared to much larger models. Would love to see more work thinking about how to improve & deploy this sort of LLM to support medical professionals
@emollick
-
Veo 3 Video Generation: Fast Results with Simple Prompts
By
–
Most of these were on the faster, cheaper Veo 3, using a dumb prompt, and worked the first time. For Ophelia, I just used the famous painting itself and the prompt: she sits up and says "Actually, I think I am over Hamlet"
-
Veo 3 Transforms Classical Art into Shakespearean Drama Visualizations
By
–
I used Veo 3 to solve Shakespearean tragedies as pictured in classical art. pic.twitter.com/k53NG24FgL
— Ethan Mollick (@emollick) 9 juillet 2025I used Veo 3 to solve Shakespearean tragedies as pictured in classical art.
-
xAI Transparency Issues on Chatbot System Changes
By
–
Everyone who is saying this is just the replybot being changed and not the system prompt for the chatbot or changes to the API – we don't know if that is true because there has been little transparency so far. xAI can address all of these concerns, but ignoring them doesn't help
-
Trust and Reliability Matter More Than Model Performance
By
–
Grok 3 was a very good model, and Grok 4 might be amazing but having a very good model is not enough – there are a lot of really good models out there. You actually want to trust the model you are building on.
-
Grok’s Reputation Issues: Why API Builders Should Care
By
–
I totally get that many Grok superfans view these as blips that no one will remember, but:
1) People remember this stuff in the real world when it is as sticky as this. I still get asked about Bard's image generation issue.
2) If you are going to build on the API, you care a lot. -
Platform Trust Critical for API Development Decisions
By
–
That is absolutely not true, especially for anyone considering using the APIs. You don't build on platforms that you don't trust.
-
Grok 3 ethics incidents raise trust concerns for Grok 4 launch
By
–
So Grok 3 has had three separate incidents where apparently unvetted changes to the deployed system caused a large-scale ethical issue and an emergency rollback. I don’t think you can do a Grok 4 launch that doesn’t at least address this honestly, if user trust matters.
-
Prompt Differentiation Risk and Third-Party Audit Solutions
By
–
If your prompt is the sole differentiator, you are in trouble. But also, why is that your problem as a buyer? And third party prompt audits can solve this.
-
Companies Must Allow Third-Party AI System Prompt Audits
By
–
At the very least, companies should allow third party audits of system prompts. You can see how much an unannounced system prompt change (or other forms of unannounced fine-tuning, who knows?) can alter behavior by just looking at what happened today on X.