$NVDA just hit YEARLY support. If you ever wanted to buy this thing, today is a good day for it.
Writing - Our thoughts on AI, software and business
Find our posts on LinkedIn. Here's an excerpt
We can fix everything.
How do you do code review in a world where LLMs produce code instantly and around the clock? You read Anthropic's Engineering blog posts and learn from an organization that increased their daily changes to the codebase more than 10x in the past 5 months. https://lnkd.in/eivaKR5B And no, a 'human-in-the-loop' is not the answer. Even though lots of humans think it is. Probably cause they don't want to feel useless next to the almighty AI. You're not in charge of reviewing code anymore. You're in charge of building systems that review code reliably and verifiably on behalf of you.
Sunday showerthought. I don't think you should build an agent into your app that your users can use. I think you should build your system (you might not even have to render html in a browser anymore) in the BYOA way: Bring Your Own Agent. Your users will bring their preferred agent(s) to interact with your system.
AI isn't here to help you do your job. You're here to help AI do your job. The sooner people understand this, the better for them.