@QuinnyPig you’d think a company that’s “doing just fine” would be able to pay more than a professors salary to a biologist if their big spooky super-anthrax model is as scary as they want us to believe. i guess fable/mythos aren’t really all that spooky!
I keep seeing large tech companies discussing the dangers of AI and AI models. They think regular people should not be able to possess AI or AI models, or be able to run them on home computers. In essence, they're too dangerous for a regular person to possess.
It is for our safety that Google, Meta, OpenAI, Anthropic, Palantir, X, etc have AI models but we don't. They are protecting us by paying them $19.99/month. If we don't let them control the AI stuff then someone could do bad things with it.
They care about protecting the people. They are definitely not doing it to make more money and collect more data from consumers. Large tech companies would not use their money to influence politicians and government to make more money and fearmonger
Quote: "This is how USAID spent your tax dollars"
I think people have lost the will to try fact-check this guy due to the volume of lies he spreads.
Also it's actually pretty difficult with this image and quite time consuming.
It just happens I had nothing better to do. /1
Will Fable 5 be nerfed when it returns?
BridgeBench has every benchmark from the original June 12 Fable 5.
The moment access is restored, we are rerunning everything.
If the returning Fable 5 scores lower than the original, you will see it here first.
Independent benchmarks exist for exactly this moment.
for someone who was VP of Research at OpenAI, Dario has a strangely poor understanding of what open source AI even means:
> in open source, you can see the source
> here you cannot see inside the model
> ultimately you have to host it on the cloud
"MyThOs iS a SuPer wEaPoN."
Anyway! Here's access for "trusted" orgs still vulnerable to CVEs with birthdays, run by execs who won't fund shit unless Gartner puts it in a quadrant, all while their Fortinet boxes are operating as nation-state APT Airbnbs.
it's 2027. you take a free-tier public Waymo to the DMV (Department of Model Variance) to do a proof-of-identity check for access to GPT 7.1.
the guy at the counter is clearly watching a Mr. Beast video in his AR glasses. "Here for that new model?" he says, barely making eye contact. he wipes his fingers on his shirt and taps at his keyboard. "Lot of you techies showing up here today." you smile politely; you're pretty sure he's just a Claude wrapper anyway.
you lean forward and stare into the retinal scanner. after a long moment, there's a soft chime. "Humanity confirmed. U.S. national. Intelligence access: Terra-class."
you sigh with quiet relief as your devices light up—notifications from a hundred agents, finally able to resume their tasks. you feel a twinge of guilt as you terminate your open-weight backup agents, but remind yourself that a joint congressional committee proved conclusively that Chinese models are non-ensouled.
you step outside and hail another Waymo. the first one passes you by. you grimace; must've burped in that one once. stupid personalized memory.
as you're waiting, your phone buzzes angrily, red notifications blaring across the screen. the Department of War just restricted access to all OpenAI models on serious national security concerns; apparently Pete Hegseth got GPT-6-Instant to say "Claude is a woman." you groan, and resign yourself to another week of merely-somewhat-superhuman intelligence.
Fable 5 is still inaccessible to the public. a twitter anon you trust says it's coming back this week. or maybe next.
Dario told the Senate in 2023 that “the scaling of open-source models” was going down “a very dangerous path.”
in other words: open weight models are fine until they get good enough to threaten the closed-model business, then they become a matter of national security.
get in losers, we're making vintage UIs w/clear visual affordances, sub-100ms performance, no unnecessary animations, bookmarkable URLs & a functioning back button