auspiv

it is very hard to take too much vitamin D

0.8% of electricity can be very critical MWs or MWhs.. rounding to 1% is 87.6 hours of electricity. could be the top 1% of prices, or grid stability, or some other grid event where having batteries able to instantly supply tens to hundreds to thousands of MW is worth a ton of $$$ and also huge increases in grid stability.

recall batteries do not generate electricity, they store it for when it is needed most.

well if there was any doubt about why openweight models were ranking lower on every iteration of the benchmarks... this should remove all doubt. to just flat out remove is very overt

starlink sats are some of the lowest orbiting satellites in space - 330-370km for v3. this helps latency as well as easy deorbit at end of life. the v3 sats have multiple 400 gbps lasers for meshing... I can't decide if your comment is sarcasm or not actually because starlink is fully functional and doesn't have any of those issues

This is a very low number for a server GPU. My V100 from 2017 is 0.9 TB/s.

A single B300 has almost 8 TB/s.

AMD MI300x from 2023 is 5.3 TB/s

Edit: my bad you are saying that this is an impressive number precisely because it is more or less normal RAM, not VRAM

someone else picked up that bit of math and has run with it and has refined it downwards multiple times. believe it is now in the range of 2^-18 or so

also consider oh-my-pi. it is much more "batteries included" than Pi

And exactly how many tokens (please do the breakdown for prefill vs decode) does a Claude $20/month plan include?

This team is in a tough spot - they'll need to innovate and get their models in a good spot quickly. They are racing the clock against various similarly sized models being burned into ASICs and running at a similar speed (or faster).

I think the question is probably more along the lines of - why not do a database with 100 TB of storage/records instead of a cache? tomato / tomato.. especially with smart caching in front of database. 100TB of flash is a good bit cheaper than 100TB of memory

I'd take qwen3.6 (3.8 as of tomorrow) 27B running at 17k per second first on the way to Sol/Fable! And then dsv4-flash-0731!

Opus 4.5 level of performance is also accessible with deepseek-v4-flash-0731 (0731 being the july 31 update) which is much, much, much smaller. 2x RTX pro 6000 blackwell can run it. 4x can run it very comfortably

Modern dell desktops (like 8th gen+) idle around 10W. Raspberry pi isn’t saving much