The fast commercialization of LLMs really hinged on a decision by Mikhail Parakhin & Microsoft to continue making Bing available after the incident where Bing/Sydney acted like a stalker to a Time’s tech reporter. In most worlds, Bing would have been pulled down & LLM deployment slowed. Now we just shrug and say “sometimes LLMs threaten you or seem to fall in love with you or whatever, you know how software is.”
Instead they fixed the issue and rolled out to prod in like two days! Legend!
Matt Fornito
2w
Lol. Well I'm sure we can recognize if OpenAI, Claude, Google, and everyone else didn't have foot full-on-the-gas "let's get to 'AGI'" (and if most these orgs hadn't fired their ethics teams) then we'd be in a different place right now.
Casey Fahey
2w
... At least we thought we knew how software is....
Sumit Dhiman
2w
It's disappointing that the new AI chief of Microsoft is using reverse gear, to make Microsoft's consumer AI push experience as degraded as possible.
Remember when OpenAI wasn’t going to even release GPT2 because it was too dangerous? Funny times.
Krista Ewing
2w
Ethan, I was listening to the Daily AI Show podcast, and this is especially relevant if you consider what they said about AI tutors and a side comment about the potential “lifetime context windows.” What could an LLM do with that, you know?
Author @ AIU | Ex-CSC
2wWas that the incident where the individual went on the 60 Minutes television news program and was proud of the fact that they “fixed the problem” in 24 hours? By making chat sessions shorter (I believe)?