@jayce_edits@GregHBurnham Agreed, things like this constantly get espoused like it's some kind of hard limit on AI capabilities, then the next training release cycle they get blown through and then theres some new thing that people cling on to as the next, for now, thing.
Bernie and AOC might actually have the dumbest idea in the world the more I think about it.
Bernie and AOC want to stop data center construction because of fears around safety, job destruction, unpopularity, etc.
But the HILARIOUS thing is that data centers will simply be built somewhere else, because AI companies WILL NOT STOP building AI capacity because it's EXTREMELY useful and is now becoming MASSIVELY profitable.
And if AI REALLY DOES destroy jobs, that job destruction would be imported into the US ANYWAY because of this magical thing called THE INTERNET, which is what carries the AI from the data centers to everyone's computers and devices.
So Bernie and AOC want to export the upside, and import the downside.
This is so unbelievably dumb on so many levels!!!!!!!!!!
@EricTopol Despite its obvious misuse during Covid, mRNA has tremendous promise for curing diseases.
Artificial RNA essentially makes curing diseases a software problem.
The machine is solving the hard problems… The machine now makes the machine. It even optimizes the machine.
Let’s stop pretending to be machines now and be humans.
Humans are amazing.
@cognition Grok 4.7 will exceed all current models.
That said, Anthropic is a great company and will probably release improved models soon.
However, the SpaceX training corpus is so awesome & unique that I would be shocked if any model is better at real-world engineering than 4.7.
An interesting effect is that models trained next year will see all the internet chatter about them making progress on incredible tasks and genuinely believe they can do it.
Until then - believe in yourself :)
astra is a powerful model and we are working to make it generally available.
we do not think it is a good strategy to keep powerful models to a chosen few.
given its cyber capabilities, we need a little big longer to do do this safely. but hopefully not too long!
Does anyone realize that if SpaceX pulls this off, Terafab would become the largest building ever constructed by floor area? More than 100 million square feet of manufacturing space, over 5x the floor area of the current largest building on Earth.
I wanted to wait before reporting because news was coming out quite rapidly after the until announcement.
SpaceX and Tesla are planning to invest roughly $16.8 billion in the initial buildout in Grimes County, Texas. It will manufacture, package and test advanced logic and memory chips under one roof because their combined demand is expected to exceed 1 terawatt of compute, more than the current global supply.
The buildout is already so massive that ASML, (the only supplier of the most advanced EUV lithography machines) is factoring Terafab demand into its own 2027 and 2028 capacity plans. Elon’s rough estimate is that 25% of its AI compute output will go toward Tesla Optimus and the other 75% toward AI spacecraft. ASML has to plan years in advance for the sheer amount of chipmaking capacity Elon wants for robots and ‘ai spacecraft’
Given some of the results I'm seeing recently, it's pretty clear Codex is a good harness.
But it will seem primitive in 2-3 months and we're about to go through another major evolution in how we use AI at the frontier. The next generation of models need more than your laptop.
It is mindblowing to me how few people realize that their lives and everything they know will change drastically in the near future.
At this point, it should be pretty clear.