Sperg Symposium
There is a difference between doing machine learning research and cancerous growth of a technology (LLM ChatBots) into every sector even those where it has little use/applications, that growth aspiring to eat half the power grid. We don't need stuff like LLMโฆ
This is really a circular claim that justifies itself. LLMs shouldn't be used anywhere because I think they're useless. This leads to LLMs actually becoming useless because they aren't used anywhere.
I really don't see any problem with companies doing A/B testing to see if LLMs can improve productivity. You literally can't prove if they're useless or not unless you test it.
Last year, LLMs were shitty at coding and maths and logic and good/okay at translation and rapid passing of long documents.
But this year, thanks to more focused training, longer context window, and better tools for the LLMs to interact with and memorize data, LLMs have gained the ability to write code and do novel maths and logic to prove unsolved conjectures (10 of erdos problems done autonomously, 11 partially solved https://www.scientificamerican.com/article/ai-uncovers-solutions-to-erdos-problems-moving-closer-to-transforming-math/)
I really don't see any problem with companies doing A/B testing to see if LLMs can improve productivity. You literally can't prove if they're useless or not unless you test it.
Last year, LLMs were shitty at coding and maths and logic and good/okay at translation and rapid passing of long documents.
But this year, thanks to more focused training, longer context window, and better tools for the LLMs to interact with and memorize data, LLMs have gained the ability to write code and do novel maths and logic to prove unsolved conjectures (10 of erdos problems done autonomously, 11 partially solved https://www.scientificamerican.com/article/ai-uncovers-solutions-to-erdos-problems-moving-closer-to-transforming-math/)
Scientific American
AI uncovers solutions to Erdลs problems, moving closer to transforming math
LLMs have recently helped find solutions to a number of minor longstanding problems. But a new plan called First Proof is really putting them to the test
๐ค1
Really Cool Memes Actually
Photo
Once gaben dies steam is gonna immediately turn to complete dogshit bookmark this
๐ข5
Forwarded from Tardigrade Giggle Chamber ๐ชฒ (Sam Hayzen)
Yeah i have a gambling problem, i cant log tf in
๐จ10โค3
This media is not supported in your browser
VIEW IN TELEGRAM
๐ญ2๐ฅฐ1๐คฎ1๐
1๐ฟ1
Really Cool Memes Actually
This is really a circular claim that justifies itself. LLMs shouldn't be used anywhere because I think they're useless. This leads to LLMs actually becoming useless because they aren't used anywhere. I really don't see any problem with companies doing A/Bโฆ
"ai solve mathematics"
Year 3 that this is claimed, niche problems are solved that mathematicians were not looking at. Apparently we need 5quadrillion datacenters not installed in capex expenditure despite the usecase being 5 theorems that they run on research PC in their research building.
Gullible idiot niggertech
Year 3 that this is claimed, niche problems are solved that mathematicians were not looking at. Apparently we need 5quadrillion datacenters not installed in capex expenditure despite the usecase being 5 theorems that they run on research PC in their research building.
Gullible idiot niggertech
๐6๐3
Thank God h didn't sell an nft during its peak or we would have had to endure 2 years of posting on how it solves artist remuneration on the internet
๐คฃ10๐ญ2๐1
Really Cool Memes Actually
"ai solve mathematics" Year 3 that this is claimed, niche problems are solved that mathematicians were not looking at. Apparently we need 5quadrillion datacenters not installed in capex expenditure despite the usecase being 5 theorems that they run on researchโฆ
Reminder that when LASER was invented, it was seen as just a curiosity and at the time was dismissed as "A solution that is looking for a problem". 60 years later and lasers are a fundamental tech in daily life and research.
โ3
And in 2002 when the dot.com bubble was happening, people cried and moaned about high dram prices and saying none of the companies and websites around them are ever gonna turn a profit and all the investment into computer infrastructure is a waste.
Truth be told, they were partially right. Most of the companies collapsed, but the companies that survived that era are now trillion dollar companies. And the internet infrastructure and copper cables that were laid out at the time were later used as the backbone for your ADSL connections for the next two decades till optic fibers started being a thing.
Truth be told, they were partially right. Most of the companies collapsed, but the companies that survived that era are now trillion dollar companies. And the internet infrastructure and copper cables that were laid out at the time were later used as the backbone for your ADSL connections for the next two decades till optic fibers started being a thing.