Robin Hanson
More precisely, the myth is of a sudden very fast leap in abilities from below or near human to superhuman abilities, during which its values radically change, & this isn't noticed or it can't be stopped. Supporting argument is: "you can't prove otherwise". https://t.co/eYcADBOAEo
The *only* reason people are hyperventilating about AI risk is the myth of the "hard take-off": the idea that the minute you turn on a super-intelligent system, humanity is doomed.
This is preposterously stupid and based on a *complete* misunderstanding of how everything works.
- Yann LeCun
tweet
Offshore
Photo
DataChazGPT 🤯 (not a bot)
Wow! I almost fell for it!

Might France ban #MidJourney as Italy did with #ChatGPT?

I wouldn't be surprised if a law is passed to punish the creation of fake impersonations made with #IA.

@EmmanuelMacron #FrenchProtests https://t.co/SAHcwGT261 https://t.co/7qTxXXgi1Y
tweet
Offshore
Photo
Paul Graham
RT @LogoArchive: KLM by Henrion Design Associates, 1961, Netherlands.

Discover Dutch more logos at https://t.co/2OFIoglbzt

#logos #branding #logodesign #graphicdesign #designinspiration #designtwitter https://t.co/YGw8co7znQ
tweet
Offshore
Video
Altryne - targum.video
RT @vocodehq: 🚨VOCODE DROP – APRIL FOOLS EDITION🚨

Prank your friends with GPT; introducing the return of

https://t.co/ruAa4Ati6w

Built with @vocodehq, Gen Z voice via @rimelabs, and of course, hosted on @Replit

happy pranking everyone! 🥳 https://t.co/0LwxiVEsXg
tweet
David Marx
In the US, internet for prisoners is often limited to over-priced email or video conferencing services. Anyone who's been away for even a year or two is going to feel like they time traveled a century when they're released into our new AI-enhanced world.
tweet
Offshore
Video
Varun Mayya
I sometimes wonder, is the singularity going to start with a demo that’s simply a tweet? I mean…would we know we’re looking at it? https://t.co/qEzV73ZNa3
Massive Update for Auto-GPT: Code Execution! 🤖💻

Auto-GPT is now able to write it's own code using #gpt4 and execute python scripts!

This allows it to recursively debug, develop and self-improve... 🤯 👇 https://t.co/GEkMb1LyxV
- Significant Gravitas
tweet
Offshore
Photo
Robert Scoble
Add “from first principles point of view” to your prompts for help designing products.

Here is the response:

Rethinking consumer electronics like a music-playing speaker from a first principles point of view involves breaking down the fundamental components and requirements… https://t.co/o4EtarGBGh https://t.co/2qJfLAtkmJ
tweet
Offshore
Photo
Stew Fortier
I think this example buried in the GPT-4 paper represents one of the more powerful ways to interact with AI.

Telling AI to do stuff is often more work than it's worth – it lacks context, writing clear instructions is hard, etc. Instead, invert it and ask the AI to instruct you. https://t.co/Gm12w0FQhn
tweet
Offshore
Photo
Yann LeCun
RT @nsjersey: This is basically a 20-year gap. Wow. https://t.co/4L37FdeVkj
tweet