I guess his business plan is contained in his company name. Safe superintelligence.
Safe AI is reasonably well defined and would be a quite different foundational model. Assuming a few things it could also be more accurate too.
So then the other LLMs are basically commodities but he has something different and better.
And he has the best pedigree to build that.
The superintelligence thing idk. But I guess to get a $30b valuation you’d have to give the investors something there and of course it would be a little speculative by nature so that’s ok.
So many problems with the existing process but questions could include
-is anyone going to take down bad quality or fake research
-is there a way to update published papers (like link rot in references )
-will the repositories be SEO optimized so people will find them
Japanese tubs are bigger partly because they are meant to have more than one person.
But in addition they are typically unit bathrooms - like prefab that are installed in a unit in like a day or two. This makes them way cheaper than USA bathrooms. In USA you need contractor, carpenter, plumber, electrician all at diff times and that makes the bathroom really expensive to install.
I assume that if bathrooms were cheaper in the USA we would have bigger/nicer ones. That could be with a bigger bath.
I've generally found they're all correct. Sometimes, I run into situations where a particular exit is closed because it's after-hours (some stations close certain exits when it's late), and Google Maps doesn't realize this, but I haven't noticed any exits being outright wrong.
I think this is probably a local issue: Google Maps is better in some places than others. Google has a huge office in Tokyo, in Shibuya overlooking Shibuya Crossing, so Tokyo is probably one of the best-supported cities in GMaps for this reason. Other cities might not have such accurate information.
Stations undergoing maintenance or upgrading may have certain exits closed or even temporary platforms that change as the project goes on. Here in Chicago, the Red Line is undergoing a major rebuild that started over 2 years ago and will not be completed until 2025. Temporary stations have been set up at multiple stops, with different entrances that have changed as the project has gone on [1].
It may have only changed a few times over the last 2 years, but it's certainly a rude awakening when you're running to catch the train and you find out your stop is now a block away from where you thought it was.
I worked for a big testing company. There are all sorts of cheating and many are easy to detect with the right tools. Maybe this is just the wrong tool.
This should have been blocked by google (possible it was like Bing?)
If you try to load ads in there are automated checks for example the url displayed must match your site url and any redirects are also cause for the ad to be rejected automatically. Phone numbers should be the same.
Also was very amusing when we shipped test boxes...that sucker ran really hot, and I got one call about the Itanium box asking for tech support help and the report was that the box was on fire.
Yes within Intel it was thought that management would not push the 960 since if did so it would be picked up by the press as validated RISC is better. But for embedded applications it was very successful, I was shipping hundreds of thousands of them per month at one point
Safe AI is reasonably well defined and would be a quite different foundational model. Assuming a few things it could also be more accurate too. So then the other LLMs are basically commodities but he has something different and better.
And he has the best pedigree to build that.
The superintelligence thing idk. But I guess to get a $30b valuation you’d have to give the investors something there and of course it would be a little speculative by nature so that’s ok.
Curious any thoughts.