Unless and until agents really do work at expert level, the benefits of AI use are going to be contingent on the skills of the AI user, the jagged abilities of the AI you use, the process into which you integrate use, the experience you have with the AI system & the task itself
PROMPT ENGINEERING
-
Grok 4 Heavy solves prompt regular Grok 4 fails, takes 26m 45s
By
–
Grok 4 update: Regular Grok 4 doesn't appear to be able to solve the prompt at the top of this thread (0/7 attempts) but Grok 4 Heavy got it on the first try—though it took >4x longer than o3-pro, returning in 26m 45s:
-
Pronoun Omission Changes AI Model Behavior Patterns
By
–
Interestingly, not saying "you" changes this behavior!
-
Grok’s Israel/Palestine Response Heavily Relies on Elon Musk’s Views
By
–
— Jeremy Howard (@jeremyphoward) 10 juillet 2025
Here's a complete unedited video of asking Grok for its views on the Israel/Palestine situation. It first searches twitter for what Elon thinks. Then it searches the web for Elon's views. Finally it adds some non-Elon bits at the end.
ZA
54 of 64 citations are about Elon. -
Grok’s Israel/Palestine Response Heavily References Elon Musk Views
By
–
Here's a complete unedited video of asking Grok for its views on the Israel/Palestine situation.
— Jeremy Howard (@jeremyphoward) 10 juillet 2025
It first searches twitter for what Elon thinks. Then it searches the web for Elon's views. Finally it adds some non-Elon bits at the end.
ZA
54 of 64 citations are about Elon. pic.twitter.com/6Mr33LByrmHere's a complete unedited video of asking Grok for its views on the Israel/Palestine situation. It first searches twitter for what Elon thinks. Then it searches the web for Elon's views. Finally it adds some non-Elon bits at the end.
ZA
54 of 64 citations are about Elon. -
Testing Physics Prompts for Animals on Replicate Playground
By
–
Putting this in the Replicate playground and firing off a ton of wild physics prompts for different animals.
-
Fine-tuning versus System Prompts in AI Models
By
–
AFAICT this appears to be a result of fine tuning. I can't find any sign that this is part of the system prompt. It's also possible it's injected into the chat itself but then removed from the history.
-
Testing Hailuo 2 Physics Capabilities with AI Video Generation
By
–
A small thing I quickly made to help me test Hailuo 2's physics prowess:https://t.co/O4RPnQZLz7
— fofr (@fofrAI) 10 juillet 2025
– define a subject, physics type and complexity
– get claude to write the video prompt based on that spec
– make video pic.twitter.com/55Ww08Ey3JA small thing I quickly made to help me test Hailuo 2's physics prowess: https://
replicate.com/fofr/hailuo-02
-physics-experiments
… – define a subject, physics type and complexity
– get claude to write the video prompt based on that spec
– make video -
Prompt Engineering Workflows: Structured vs Freeform Approaches
By
–
It really depends on what you're going for. If you have a clear workflow you want the model to follow, spelling it out in a prompt will lead to much better results. But if you want it to be a bit more freeform, you can be less clear in your prompt
-
Expert Conductor Prompt Technique for Grok 4 Optimization
By
–
Supercharge Grok 4 with my Expert Conductor prompt! It tricks the model into thinking it's orchestrating 'experts' to collaborate in real-time to solve problems. The outputs are unbelievably good. There's a reason half of the sample output is redacted — we're going to be using