"they're only withholding the model for safety as a marketing ploy" is such a dumb take and has been for most of a decade. you can think they're wrong about ai risk but nobody is running gigabrain plans to forgo enormous certain profits now for.. theoretical secret future profit
Users disagree on AI labs' motives around safety and releases, with some defending decisions like GPT-2 weights for raising awareness while others view actions as marketing to claim priority.
"they're only withholding the model for safety as a marketing ploy" is such a dumb take and has been for most of a decade. you can think they're wrong about ai risk but nobody is running gigabrain plans to forgo enormous certain profits now for theoretical future profit
also laughing at the gpt2 weights decision is dumb too. wrong move in hindsight, but wasn't clear at the time. we were surprised by the model and didn't know the forward learning curve. the web sloppification could have taken 6mo not 6yrs, giving the world a heads up was good
we're obviously at a critical time in the ai race, everyone involved presumably believes present advantages is worth more than marginal someday advantages by a lot, all the incentives point greedily towards maximizing things today
we're obviously at a critical time in the ai race, everyone involved presumably believes present advantages is worth more than marginal someday advantages by a lot, all the incentives point greedily towards maximizing things today
"they're only withholding the model for safety as a marketing ploy" is such a dumb take and has been for most of a decade. you can think they're wrong about ai risk but nobody is running gigabrain plans to forgo enormous certain profits now for theoretical future profit
@nickcammarata Not sure I follow. Surely claiming your model is extremely dangerous could be marketing of utility? I don't think that's what's happening, but you can't rule out profit motive atm. @jackclarkSF has been very consistent though, which is why I take Anthropic more seriously.
They don’t have the compute to roll it out widely, but also needed to make sure everyone knew they got there first.
OpenAI will roll out something better than mythos by end of month and anyone will be able to use it.
The two biggest AI labs are racing to go public lol, it’s ALL marketing.
"they're only withholding the model for safety as a marketing ploy" is such a dumb take and has been for most of a decade. you can think they're wrong about ai risk but nobody is running gigabrain plans to forgo enormous certain profits now for.. theoretical secret future profit