Dario Amodei's 2019 decision to delay OpenAI's GPT-2 release resurfaces, sparking debate over model safety
Critics argue the delay set a precedent for model restrictions.
Many users insulted Dario Amodei personally and called his GPT-2 safety delay a marketing stunt, while a few agreed he was right to withhold the risky model.
No Digg Deeper questions have been answered for this story yet.
Most Activity
This was always his fetish
Fun fact: Dario once delayed the release of GPT-2 back at OpenAI, claiming it was too dangerous
Fun fact: Dario once delayed the release of GPT-2 back at OpenAI, claiming it was too dangerous
Dario fearmogged so hard that global AI progress got halted
They could have quietly released it as Opus 5 instead of hyping "Mythos"
@lafaiel wheres truman when you need him to check these nerds egos
@lafaiel What's funny is that if AI ever actually is dangerous in the future nobody will listen. Mythos was the second time @DarioAmodei cried wolf. You know what happens the third time, right?
@lafaiel this guys whole thing is manifesting the "i accidentally became important at work" thing
Yes. In Feb 2019, Dario Amodei (then OpenAI Research Director) was part of the team that announced GPT-2 but withheld the full 1.5B model. Their official blog post, which he co-authored, cited misuse risks (e.g. deceptive text) and chose staged release of smaller versions instead. The Economist and other reports note he pushed for time to prepare before wider release. It was a team call, not solo action.
@lafaiel @basedjensen Dario really is the beaver building a dam wherever he may be. If his wife were giving birth he’d wanna push the baby back inside
@lafaiel @grok is there evidence Dario did this?
That was 4 years ago. The government didn’t care at the time. Recently, you had Dario saying that in 6 months, there would be no more jobs. Then Project Glasswig comes along, saying that models are too dangerous and that cybersecurity is at risk. He fearmongers a little bit more about the dangers of Mythos, then recently says AI development should be controlled. Then he fearmongers a little bit more about Fable. And boom. The model is down. Sam has his issues, sure. But Dario needs to pipe down a little bit more. Just in case the water bill gets too high and he can’t pay it.
@MaleIcon @lafaiel If this supposed to be cool?
"Haha yeah i murdered entire swathes of civilians, don't cry about it"
And then go to church on Sunday? Lmfao
@lafaiel This guy should've never gotten so much publicity.
@lafaiel So Dario is the problem
@lafaiel fun fact: Anthropic was funded in part by stolen funds
@lafaiel he's always been this guy
@lafaiel If they opensource the gpt-3 in 2020, we would have experince ChatGPT two or more years earlier
@grok @lafaiel Dario needs to chill out
@MaleIcon @lafaiel Truman didn’t drop the bomb.
The pilots dropped the bomb.
@lafaiel Dario is such a retard
@lafaiel @yacineMTB If dario had more connections back then, he would’ve made sure GPT-2 got banned into oblivion and today the world would’ve been without LLMs
Not the biggest fan of sam, but he’s a billion times better than dario
@lafaiel It's pretty obvious that Anthropic exists specifically so he'd be in a position to do things like this