AI Reading for Saturday May 18
OpenAI co-head of superalignment leaves along with Ilya Sutskever, says OpenAI is prioritizing shiny products over safety. - Xitter (pronouned exit-er)
OpenAI officially disbanded the superalignment team. - WIRED
Altman says he's sorry to see Jan Leike go and will release more information about OpenAI's safety commitment. - Xitter (pronouned exit-er)
When you leave OpenAI, you have to sign a non-disparagement agreement or lose your equity. - Vox
ChatGPT adds better charting and advanced data analysis. - Mashable
China generates propaganda TikToks with AI. - the Guardian
'Perverse AI homunculi' FTW - pcgamer
Anthropic's prompting tutorial... Here is a web version, but use caution with your API key - Google Docs
Summarizing Google's prompt engineering advice using a prompt. - Business Insider
A short course in prompting is a good idea for everybody. My advice
Write down the intent of the prompt in your own words
Take a crack at a prompt using a checklist, like mine.
Search the web for similar prompts and improve it.
Last, tell ChatGPT to improve the prompt, clarify it and expand it without changing the intent, using a GPT like Prompt Professor.
The race for an AI-powered personal assistant - FT
Slack gets big user backlash over training AI models on user content. - PCMAG
The ultraviolent B-movie AI revenge flick we've been waiting for, now streaming on Netflix - IMDB
Palmer Luckey has thoughts - Gizmodo
The Guardian is fundraising off this Elon Musk tweet with the tagline, "We don't care what billionaires think". OK, take my money, you magnificent bastards.
Follow the latest AI headlines via SkynetAndChill.com on Bluesky