People forget that AI models are great prediction systems, but as such the make very interesting blunders. If you are generating an essay, summarizing a meeting or just making cat memes that level of precision is acceptable. But I don’t see AI replacing human in anything that requires exactitude, predictability or intuition.
I really like microbin to copy paste files around.
And who will be in charge of defining “Best Behavior”?
If I remember correctly the agreement is some kind of lease, so the city/county doesn’t really own them after they are installed.
An GMKtec EVO-X2 with 128GB was in my radar earlier last year, but at $3200 this is not something that I can justify buying.
Do you mean in addition to Win11?
9GW should be the compute load goal, to which you need to add the mechanical and administrative loads. At higher scales they gain significant efficiencies which translates to market advantages.
Not that it would matter for this conversation, but at hyperscalers levels, the energy required for mechanical loads is under 20% of the compute load. Wouldn’t surprise me if ~10% can be achieved at multi GW scale. Thus about 11GW total energy.
Kevin is a POS and I really doubt this DataCenter will ever be built. Having said that, this article is also full of fearmongering trash.
”Turn out, it could create a massive heat island capable of devastating the area’s ecology, said Robert Davies, a physics professor at Utah State University.”
Wow. This is absolutely has to be the reason. There is nothing better than a recurrent revenue stream. Look at Spotify, Netflix business model.
Me too, I was waiting for a good deal in a GMKtec EVO-X2, but that is not going to happen.
Did you use ChatGPT to write this?
I believe seat belts were invented in the sixties and mandatory some time in the eighties. Unfortunately safety lags innovation.
It works great. I print mostly in PETG so bed temperature might be a factor.
I clean mine with isopropyl alcohol every other print or so.
Having a passport card will definitely make it easier for the deportation official to release you after assault you, put you in handcuffs and take you to the station for speaking Spanish in a Mexican restaurant.
This is too big for a bike, but I self host Traccar, and installed Sinotrack st901 in my cars years ago. SIM card cost is about 50 cents a month each.
Edit: check Mictrak MT710
I’m paying for a search engine “kagi” just because Google search results have more advertising than time square. It kind of sucks since we got used to search for free but at least I can get relevant results rather than advertising when I search.
I have been trying to make sense of all AI Capex announcements for a while and I don’t get it. So please help me out if you know the answer.
US Investment in 2024 ~$400b, 2025 ~$500b, 2026 ~600. Global investment 2026 1.5~$2.2t. Let’s say $2t US Investment by end of 2026. Investment will continue into the future but let’s assume that is not the case. Also that GPU will be obsolete in 5y. So, they have 60m to recoup $2t +ROI ~10%. So about $40b a month, US has a labor force of 170m. Thus, AI industry needs the equivalent of ~$240 per month per each employee. I don’t see myself or my employer paying this for AI any time soon.
And China’s moat is the ecosystem that they are building around EVs. The mining and refining tech for lithium, battery tech, motors, drivers and manufacturing that will be impossible for any other country to replicate.