As a diver, if you jump into moderately warm water (the Great Barrier Reef for instance), that won't save you either. Mine overheats and dies after 15-20 minutes. It took diving near Seattle to keep it going continuously.
You might have to mess around with the software to figure out the details, but from a hardware perspective a hall effect keyboard should be at least able to infer the speed a key is travelling at when bottoms out.
Hey, Replit employee here. I'm pretty sure this isn't us (definitely isn't our marketing team's MO AFAIK). Can you email me some examples at james @ replit dot com so I can look into this?
GST and VAT are the same thing, different countries just call them different things.
As an example, GST in Australia is in the price on the shelf. You can also claim back the GST you paid on your input goods, which means you effectively pay the difference on the GST your customers pay you, and you paid your suppliers (which is the value added!).
FWIW, I had the same issue with my 3090 (though I believe that uses a slightly different port?). I was using a custom cable like this guy. Nvidia replaced it under warranty, and I went back to using the (ugly) provided adapter.
TLDR: If he buys a car a year, and drives the same number of miles as the average American, his total emissions are probably equal to someone who drives 28k miles a year.
Carbon emissions for a Model 3 vs a Toyota Corolla even out after 13'500 miles according to Argonne National Laboratory [1], which is slightly less than the average an American drives per year (14'263 miles [2]). Assuming that he drives as much as the average driver, his cars generate as much Co2 as a Model 3 (definitely not true for the Cybertruck, but he probably has low-build-emission ICE cars in the other 8 to lower the average), and he buys a car a year, he has roughly the equivalent emissions of someone who drives twice the average number of miles each year. For reference, a long haul driver (of which there are 300k-500k in the US [3]) drives 100-110k miles [4] a year (7-8x the average).
A lot of analysis is done using CSVs being pushed in and out of Excel. Doing so strips the formatting. Please understand the workflows being crying “skill issue”.
The model they use (BAAI/bge-small-en-v1.5) produces embeddings 384 wide, at float32 that equals 1536 bytes each. The size of all the vectors for 100 items is 153kb. Calculating the dot product of that against a query will be under measured in nano seconds, even with a naive implementation.
This is the Australian Broadcasting Corporation's website (the Australian national broadcaster a la the BBC). As a whole coverage is pretty good, but the website has gotten more clickbaity in recent years.
We actually announced that Ghostwriter is now free for all users last week: https://blog.replit.com/ai4all. The free tier is using a smaller model for chat than the paid version for cost reasons, but it has all the same integrations (inline suggestions/code transformation/etc).
(Disclaimer, I work at Replit on the AI team, but don't represent the company)
That isn't what is being described here. They are just providing additional context to ChatGPT using its plugin API. It's still trained on large amounts of public text data.
I'd say it's more "doesn't bother to tread" rather than "doesn't dare to tread". There probably is not a lot below 1.5km that would be of much interest to the US Navy.
GPTs are not trained on themselves, and what little knowledge they have of their own capabilities is limited to the discussion of it in their training set. This is a hallucination.