That is just one of the major underlying problems IMO. If ChatGPT et. al. were charged out at a modest profit, who would pay for these services that most people who are using LLMs non-industrially get for free at the moment? Watch an hour of ads to get one animated gif or the answer/hallucination to some trivia? $300pm Basic Tier with reduced ads, $1000pm no ads?
It smacks of the old: we lose on every sale but we are going to make it up in volume! It is not like software where the development costs are amortised across future sales and hosting is cheap as an ongoing cost - if queries average 36c that equates to ~2kWh or >7MJ of energy every time someone hits the button. 18 queries and that is a gallon of petrol...
They have managed to get enough VC funding that the heap of dollar bills has started to compost and generate heat.
As others have posted, it does look like the AGR from the 1960s only maybe now with AI and blockchain?
What gets me is that this is critical national infrastructure - it cannot fail through one fault, or have the redundancy located such that a single failure can take that out at the same time, like backing up your data to another partition on the same hard drive.
The terminals that went dark AFAIK use power in the 20-30MW range which you can fit on a standby basis in a handful of shipping containers onsite at a cost that would be a rounding error compared with BAA turnover and profits. I can see a new 3MW one for sale for less than $100k which is peanuts even when you take into account the integration costs.
A typical datacenter uses this kind of power and I bet their redundancy is a lot better.
Very early 80s. Computer room was unlocked at 8am, so if you skipped breakfast you could get a whole hour in until lessons started at 9.
I remember the spring mornings, the cursor on the Commodore PET fading up as the CRT warmed and the feeling of anticipation before the first line of BASIC or assembly went in using the still-cool keyboard. With 8KB of RAM (32 on the special machine) the possibilities were limitless...
Ultimately, is what Bill Gates is doing on the whole beneficial to mankind? Would we be better off if he spent his money on superyachts or space rides?
What he appears to be about is applying a ruthless business brain to the business of charity, which is absolutely necessary to provide focus and get the money to where it is most needed, be it research, production, distribution or even education.
Most governments seem to be incapable of making long-term benefit-of-all investments and decisions and it seems to be left to philanthropists or even more commercial interests to do this. In that light it looks like Mr Gates is doing more good than bad...
This must affect, well, the 10 people who bought the thing...
I can assure you having been in Kolkata and Dhaka during the monsoon, glaciers are are minor part of the equation. In Meghalaya State they measure the annual rainfall in metres, not cm.
Like food, there is enough water for everyone if it is distributed fairly and used economically.
I think it is a clue that, at the moment, AI implementations are not living up to the hype.
In theory it looks fine: bit of speech recognition, very limited menu options, not much variation, no need to interpret anything other than food and drink and humans are in the loop as a sanity check because they have to prepare and serve the items. Around the level of a Uni project, really.
The fact that it appears to be unsatisfactory in some respects (and joke about IBM but this is not the most demanding thing they have ever done) does make you wonder about practical applications of AI-with-everything...
The "cutting edge" is getting rather dull. -- Andy Purshottam