AI Cheapness Lie, being pushed hard now by big tech
(1) Spend hundreds of millions of dollars on training competitive foundation models
(2) Convince everyone that you’ve found some great tech innovation that now allows anyone to do the same, training their own competitive foundation models at home for $10, and so there’s never a point for anyone to spend big money on training their own foundation models again
(3) This kills investor appetite in investing in training of competing models
(4) By the time they realize the lie, it may be too late, monopolization success.
Why do you think OpenAI is spending nearly half a billion $$$,$$$,$$$,$$$ dollars on training GPT-5 right now, as we speak, instead of just $10?
Are they just stupid? No.
AI Cheapness Lie.
(1) Spend hundreds of millions of dollars on training competitive foundation models
(2) Convince everyone that you’ve found some great tech innovation that now allows anyone to do the same, training their own competitive foundation models at home for $10, and so there’s never a point for anyone to spend big money on training their own foundation models again
(3) This kills investor appetite in investing in training of competing models
(4) By the time they realize the lie, it may be too late, monopolization success.
Why do you think OpenAI is spending nearly half a billion $$$,$$$,$$$,$$$ dollars on training GPT-5 right now, as we speak, instead of just $10?
Are they just stupid? No.
AI Cheapness Lie.
👍21🔥5💯5❤3
OpenAI Paper: Model evaluation for extreme risks
“Model evaluation for extreme risks should be a priority area for AI safety and governance. There
are many challenges ahead for finding effective evaluations and building governance regimes that incorporate them; we encourage further work in this area.”
Translation: Inside look at the censoring of GPT-4.
Bonus: First glimpses at the governance regimes who will soon unilaterally rule over us.
Paper
“Model evaluation for extreme risks should be a priority area for AI safety and governance. There
are many challenges ahead for finding effective evaluations and building governance regimes that incorporate them; we encourage further work in this area.”
Translation: Inside look at the censoring of GPT-4.
Bonus: First glimpses at the governance regimes who will soon unilaterally rule over us.
Paper
🤬32❤9👍9🙏1🙊1👾1
This media is not supported in your browser
VIEW IN TELEGRAM
Sam Altman on Open Sourcing GPT-6
😐15🤬9🤔4👍2❤1👌1