Claude is retarded. All of these models are.
I wanted to schedule a skill every day at 8:00 am. Claude decided to schedule it at 7:57 am "to avoid the on-the-dot" surge.
I SAID 8:00 AM! Do the darn thing the way I asked you to do it!
You gotta be crazy to trust these models.
Don't just buy GPUs for finetuning LLMs.
NVIDIA has built libraries for *drop-in* GPU acceleration of common ML tasks:
• cuPy for NumPy
• cuDF for Pandas
• cuML for Scikit-Learn
• cuGraph for NetworkX
Rapid iteration is key. Use your GPU for more than training.
We can design AI systems to be both super-intelligent *and* submissive to humans.
I always wonder why people just assume that intelligent entities will necessarily want to dominate.
That's just plain false, even within the human species.
Interview tasks are going to change massively. Whip up a 500 word paragraph on [INSERT TOPIC] using ChatGPT. The differentiating points amongst candidates will naturally be prompt engineering.
Math language is the supreme pedantry. We take a word with a well-established meaning, give it some arbitrary super-precise definition that is only applicable in very narrow domains, and then insist that guy on TV is stupid because he used the word "continuous" wrong.