Here’s what amazes me about the reaction to LLMs: they were designed to solve NLP, stunningly did so, and then immediately everyone started asking why they can’t do math well or reason like a human being.

▲

Peritract 4 days ago | parent | next [-]

LLMs were pitched as 'genuinely intelligent' rather than 'solving NLP'.

We had countless breathless articles about free will at the time, and though this has now decreased, the discourse is still warped by claims of 'PhD-level intelligence'.

The backlash isn't against LLMs, it's against lies.

▲

aNoob7000 4 days ago | parent | prev | next [-]

Because the heads of tech companies jumped on TV and said that AGI was around the corner to basically prepare for job losses.

They just can't shut up about how AI is going to either save us all or kill us all.

	▲	joquarky 3 days ago \| parent \| next [-]
		Well the job losses have certainly arrived. Whether that is due to AI, WFH->offshoring, or end of ZIRP is anybody's guess. All I know is any tech meetups I go to are full of people looking for work and the recruiters that normally stop by have vanished.
	▲	lukev 4 days ago \| parent \| prev [-]
		The VC economy depends on a hype cycle. If one doesn't exist, they'd manufacture one (see web 3.0), but LLMs were perfect.

▲

nielsbot 4 days ago | parent | prev [-]

maybe a classic case of the sales team selling features you haven't built yet