I just published a new issue of Beyond the Prompt:
Before You Trust the Output
AI can sound confident even when the context is weak.
Before trusting the answer, ask where it came from and whether it can be checked.
https://t.co/w0WRjcdKIv
This works really well btw, at the end of your query ask your LLM to "structure your response as HTML", then view the generated file in your browser. I've also had some success asking the LLM to present its output as slideshows, etc.
More generally, imo audio is the human-preferred input to AIs but vision (images/animations/video) is the preferred output from them. Around a ~third of our brains are a massively parallel processor dedicated to vision, it is the 10-lane superhighway of information into brain. As AI improves, I think we'll see a progression that takes advantage:
1) raw text (hard/effortful to read)
2) markdown (bold, italic, headings, tables, a bit easier on the eyes) <-- current default
3) HTML (still procedural with underlying code, but a lot more flexibility on the graphics, layout, even interactivity) <-- early but forming new good default
...4,5,6,...
n) interactive neural videos/simulations
Imo the extrapolation (though the technology doesn't exist just yet) ends in some kind of interactive videos generated directly by a diffusion neural net. Many open questions as to how exact/procedural "Software 1.0" artifacts (e.g. interactive simulations) may be woven together with neural artifacts (diffusion grids), but generally something in the direction of the recently viral https://t.co/z21CP5iQfu
There are also improvements necessary and pending at the input. Audio nor text nor video alone are not enough, e.g. I feel a need to point/gesture to things on the screen, similar to all the things you would do with a person physically next to you and your computer screen.
TLDR The input/output mind meld between humans and AIs is ongoing and there is a lot of work to do and significant progress to be made, way before jumping all the way into neuralink-esque BCIs and all that. For what's worth exploring at the current stage, hot tip try ask for HTML.
this is the MOST important 4 minutes you’ll watch on AI this year.
anthropic built a model so good at finding vulnerabilities they didn’t release it to the public.
>CLAUDE MYTHOS PREVIEW
it’s unreleased to the public and here’s what it did in a few weeks:
>found a 27-year-old vulnerability in OpenBSD
>caught a 16-year-old flaw in FFmpeg that automated tools missed after 5 million tests
>chained together multiple linux kernel exploits autonomously. no human steering.
AWS, google, microsoft, apple, nvidia, crowdstrike, JPMorgan. all got access.
Anthropic committed $100M in credits to let these companies hunt vulnerabilities in their own systems before attackers do.
>93.9% on SWE-bench verified. >77.8% on SWE-bench pro.
nothing else is comes remotely close. Anthropic just pulled away in this AI race…
Just built a website in minutes with @replit AI
→ Prompt → Files → GitHub → Hosted free on GitHub Pages.
Even crazier: when something broke, I told the AI — it took a screenshot, showed the issue & fixed it.
Props to @amasad#AI#WebDevelopment
Just started using Cursor and wow. With GPT-5 built in and solid docs, building full apps is easier than ever. Solving hard problems takes minutes, not hours. The life of a programmer is becoming more fun for sure.
#AI#CursorIDE#GPT5
@DThompsonDev As you moving up to the latter you are still learning. I’ve encountered people that brings you down and people that brings you up. But the most hurtful are the people that you think can help you while they are hurting and bring you down.