Advancing charting technology through cutting-edge engineering.
Joined June 2016
- Tweets1.2K
- Following137
- Followers103
- Likes3K
Thanos retweeted
the pH of the vagina is so acidic it can completely dissolve $1.25T in market capitalization
Thanos retweeted
JUST IN: Solana records $84,000,000 in tokenized $NVDA stock volume, more than all other chains combined.
Thanos retweeted
Couple days in
๐ Thessaloniki ๐ฌ๐ท
Walked for about 3 hours yesterday through main areas. Sunny & nice for early Nov.
Quite impressed. Way more chill than Athens.
Clean. Less graffiti slop. Nice green areas.
Very affordable. Lot of young people around, looked good, healthy.
Guess tons of beach / day trips nearby. International airport.
Could be solid spot to camp out !
Thanos retweeted
Western media : trump is losing elections blah blah
Middle east media : tensions in the region blah blah
European media : Russia is a threat blah blah
Russian media : MASSIVE SNOWBALL IN RIVER
Thanos retweeted
The much less than 2.1 children/woman (less than 1 children/woman in South Korea) is the biggest problem humanity has currently. People are lazy in thinking and just wrong in explanations, mixing their favorite political themes to this issue. No country is attempting to honestly understand it, let alone to fix it or mitigate it.
youtube.com/watch?v=VQqelrmxโฆ
Thanos retweeted
I trained a 3M-parameter model to write JPEG XL art. This image is one of its programs, 29 bytes.
It learned from random programs and 1,442 of my left/right clicks on its images. It never saw art made by people.
Write-up and an in-browser blind test: nafty.is-a.dev/posts/jxl-artโฆ
ALT Neon Chevron: glowing red, blue and cyan diagonal light trails on black that rise from the left and fold downward at a sharp tip.
Thanos retweeted
Which group of these four is most likely to drive for islamization of Europe:
0%Buddhists
0%Christians
62%Jews
38%Muslims
13 votes โข Final results
Thanos retweeted
I Agree with Sam Altman.
First of all, local models are useless, they're dumb and slow
You'll spend ~$1000 on 128 GB RAM to run a local model
And you won't get anything done with it, dumb models at 8tps are useless
A $20 subscription will get more things done
And for 95% of the time, RAM is not being used so it's doing less work there compared to in a data centre where utility is more
Readers added context they thought people might want to know
Local models on suitable 128GB hardware often run at 30+ tokens/second and offer practical uses; 128GB DDR5 RAM kits currently start above $2200, not ~$1000.
datahardware.ai/tools/benchmarโฆ
theramageddon.com/128gb-ddr5-ramโฆ
