-
Posts
3,838 -
Joined
-
Last visited
Everything posted by danny.
-
You're describing many posters on here, not just Tommy and Kenny
-
Do you genuinely believe this?
-
Yea so nice. Just whack the climate control on 15 mins before you get back in, game changer. Same in the winter too.
-
Bold of you to assume there is an end
-
This is quite shocking to read. A fried beef tomato half is amazing in a fry up.
-
Are you including a fry up in this?
-
It doesn’t matter what people think, though, the average person has no power. They don’t get to choose the direction of global affairs. Those choices are made for us and we have, every now and then, the illusion of democracy.
-
The latter
-
Just to remind you again, we don’t actually get to choose which.
-
I don't understand what AI has to do with the "idea of supremacy".
-
It's genuinely impressive how you make every topic about this.
-
Fair enough, we are talking about different things then. I agree a general purpose data centre isn't ideal in space. A self contained one that that mainly holds LLM models and doesn't need to talk to other centres loads wouldn't have an issue with latency though, obviously some parts would need syncing like billing/usage/limits etc but thats not as time critical as requests. The main issue is cooling, they would need an insane amount of radiator area to get rid of the heat via infrared. And also, sorry, not V2 - I had mistakenly thought StarLink was on V1 now - it's already V2. The next gen (which I guess is V3) is going to be at a lower orbit and SpaceX have claimed < 5ms latency. You're right, current is closer to 25-50 most clients (which again would be fine for an LLM only data centre, and TBH probably a lot of general hosting, it wouldn't be perceivably slower for a lot of less urgent stuff like general website hosting, esp once you take browser and CDN caching into account)
-
I said V2. Not sure why you’ve pivoted from LLM API calls to real time bidding.
-
Latency isn't an issue at all. APIs are nowhere near instant now, even the quickest models have perceivable delays of seconds not ms. Starlink V2 latency is < 5ms, assuming AI satellites are in that ballpark or even double that - it wouldn't be noticeable for gaming let along LLM API calls.
-
Which countries are pushing to govern this?
-
It would be pretty easy to train a model on that data to bypass it - then it's an arms race that eventually would be won by AI as it would merge into non-AI footage.
-
it's also going to make feasible denial for many crimes an absolute minefield.
-
Imagine getting a train to Leicester to watch a play and your first experience is Granby Street
-
Oh sorry, I was too busy on my far right marches to respond to your post necromancy.
-
Huge numbers of more landlords buying more properties to rent has caused massive demand. OK...
-
Good stuff, look forwards to everything being affordable again and there being plenty of rental properties available at cheaper prices for the people who need to/want to rent. No idea what happy clappers means, people are happy Zach has fixed everything and are clapping the move? Good to know it's all in hand.
-
Yes, because costs have gone up, mortgage interest rates, far higher maintenance costs (esp. post-COVID), and higher taxes plus legislation that makes it more expensive to rent out. And of course supply/demand adds to it too. I'm sure they will all come back down once Zach has put the private landlords in jail though. Just have to be patient. And Blackrock definitely won't want to make as much money as they possibly can.
-
