Modern grids favour flexibility over fixed baseload generation (like nuclear) though. When you turn off a nuclear power plant its operating costs basically stay the same, which is horrible when you could cover your whole consumption with basically free solar/wind.
actually nuclear is terrible in a grid increasingly full of nearly-free variable sources (solar&wind). The nukes need to stay at 100% all the time selling their power at a high fixed price to have any remote chance of being economical. Cheap variables push nuke's expensive power off the grid during the day, and increasingly into the evenings with batteries. This is deadly to the economics of nuclear.
AI is not cheap to run no matter where it is running. The price we get charged today for AI is a loss-leader. The actual cost is much higher, so much higher that the average paying user today would balk at what it actually costs to run. These AI companies are trying to get people hooked on their product, to get it integrated into every business and workflow that they can, then start raising prices.
Even if you live somewhere where it does, that is not remotely "almost free", and lots of places the payback period is more in the range of 10-15 years even with subsidies.
Well, when a companies have 100billion dollar incentives to make discoveries like this, I don't know if we should assume this is the only optimization that will happen.
Given that increasing model size doesn't yield proportional increases in intelligence, there is a world where these datacenters don't have a positive ROI if we make these models even a fraction as effective as the human brain.
I think that either investors were extremely skittish that the stocks might crash and jumped at the first sign of trouble (creating a self-fulfilling prophecy) or they were trading on non-public information and analysts who don't have access to said information are reading too much into the temporal coincidence of the Google Research blog highlighting this paper.
Well considering basically the entire market was down these past few days, Google included, its unlikely attributable to this paper alone. Its most likely correlated with general war/trade route restrictions/potential recession fears, or at least, more correlated with those than it is with this paper.
This paper was released a year ago and was probably part of how google got to 1m context before other labs.
>Instead, we should make it illegal to discriminate based on criminal conviction history
Absolutely not. I'm not saying every crime should disqualify you from every job but convictions are really a government officialized account of your behavior. Knowing a person has trouble controlling their impulses leading to aggrevated assault or something very much tells you they won't be good for certain roles. As a business you are liable for what your employees do it's in both your interests and your customers interests not to create dangerous situations.
My point was only that you may not have checked but you know about the 737 Max. Do you know about software failures from Google, Microsoft, OpenAI, etc. killing someone? They certainly have but it doesn't get the same press.
Do you honestly expect us to just turn a blind eye to Russian assets spreading disinformation in a time when Russia is literally waging wars of genocide in Europe? No. Strip his nationality and let him enjoy his Russian passport.
It's hard to put into words, but you're eroding the social contract through your actions. People with conditions get accused of faking it all the time, and it sounds like you're actually faking it.
If he was doing that to get faster treatment at a hospital or even just a restaurant or something then I'd agree. But by doing it to get faster treatment at the TSA check he's literally doing everyone else a favour.
The argument is that if tricks like this were to become widespread, they may start requiring certified medical documentation (or other hurdles) for said faster treatment, making life even more annoying for people with genuine issues.
In that case would actually increase security, right? Ans with genuine medical issues it should be no problem to get the necessary documentation. Either way, the consumers win.
If they opted for a pat down for 6 years, then faster treatment clearly wasn’t the goal. Metal detector + swabbing is not faster than the scanner either.
Depends on the trip. I've timed many trips bike vs. car in my city. Bike is usually faster (or very similar time) because the average speed through a city is actually pretty slow. There's a lot of "hurry up and wait" with cars (rush to get to the next red light) and on my bike I'm frequently passing long lines of cars stopped at lights.
And parking is a time sink. There's a place in my city that has huge parking garages with lots of parking but you still have to drive through a few levels of the garage, park, and then walk back down a few flights of stairs, then walk to your destination. I just park my bike right outside of my destination with the wheel lock. Street parking is always awful in populated cities, and I never have to worry about it. I always park right at my destination, where ever it is.
In suburbia, cars are faster because the average distance per trip is a lot longer. But it's ironic that the reason why the average distance is longer is BECAUSE it was built for cars so everything gets spread out! Cars are a solution to a problem that they created.
Not the case as density increases: once roads reach their capacity, space inefficiency quickly becomes time inefficiency. That's why some cities have started introducing congestion pricing.
A healthy power network will have a variety of generations sources available.
reply