Hacker News new | past | comments | ask | show | jobs | submit login
Twitter reacts to HN not appreciating LLMs (twitter.com/8teapi)
9 points by spacevacuum on April 7, 2023 | hide | past | favorite | 3 comments



Good. I have your attention.

I also guarantee that these AI bros do not totally trust the output of an LLM like ChatGPT or GPT-4 for it to be reliable in anything involving very high risk situations such as medical, financial and legal advice without a team of more humans spending more time reviewing its nonsense.

Perhaps to the AI bros, the truth hurts and it has certainly shocked them in their own hype bubble, since LLMs not only are black-boxes they lack of any transparent reasoning and explanation of their own decisions; a fundamental chronic issue which affects deep neural networks which they are all using.

The only serious safe use case is summarization of existing text, hence the abundance of apps, feature(s) and integrations all over the place for this. Eventually that use-case will soon be as copyable as a flashlight app and everything else in LLMs is another gimmick.

Once again, this so-called 'revolution' into LLMs + chatbots is a total grift and is only giving ideas to the incumbents of the AI race which is Microsoft, O̶p̶e̶n̶AI.com, and Google.


No better sign the hype-train has gone into runaway acceleration than a bunch of social-media clowns running around feeling superior because "WE'RE EARLY BRAH", the mating call of the dumbfuck.

See also crypto, NFTs, the metaverse, etc...


There is no hive mind “Twitter” or “HN.” Just one twitter user reacting to a post by another twitter user about a single HN commenter’s take.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: