Sorry, Sam and Elon, we have not reached the Singularity
garymarcus.substack.com
Sorry, Sam and Elon, we have not reached the Singularity
1–8 of 8 posts
Re: Sorry, Sam and Elon, we have not reached the Singularity
#2Because that’s the definition of a singularity. Unless they can point ant any concrete examples, the claims are completely meaningless.
Re: Sorry, Sam and Elon, we have not reached the Singularity
#3Re: Sorry, Sam and Elon, we have not reached the Singularity
#4Re: Sorry, Sam and Elon, we have not reached the Singularity
#5Exactly what technical advancement in the field of LLMs have LLMs themselves contributed to? Because that’s the definition of a singularity. Unless they can point ant any concrete examples, the claims are completely meaningless.
Re: Sorry, Sam and Elon, we have not reached the Singularity
#6Exactly what technical advancement in the field of LLMs have LLMs themselves contributed to? Because that’s the definition of a singularity. Unless they can point ant any concrete examples, the claims are completely meaningless.
Likely more than we realize.
Occams razor suggest there is nothing.
Re: Sorry, Sam and Elon, we have not reached the Singularity
#7Earlier quoted context omitted.
Likely more than we realize.
You mean anthropic and the others wouldn’t hype that up like there was no tomorrow? Occams razor suggest there is nothing .
Younperfeftky know that basically all software Deva use chatgpt or the like constantly to answer questions.
In that second some of the progress was already attributed to llms.
Both openAI, anthropic and all others guaranteed use llms to both suggest, implement, and orchestrate search for new model archetectures and algorithms.
Re: Sorry, Sam and Elon, we have not reached the Singularity
#8Earlier quoted context omitted.
You mean anthropic and the others wouldn’t hype that up like there was no tomorrow? Occams razor suggest there is nothing .
This is just stupid. Younperfeftky know that basically all software Deva use chatgpt or the like constantly to answer questions. In that second some of the progress was already attributed to llms. Both openAI, anthropic and all others guaranteed use llms to both suggest, implement, and orchestrate search for new model archetectures and algorithms.
The most successful example so far of en llm developing on its own is the C compiler that ended up a broken buggy mess.