Tootfinder

"Because sometimes the most sophisticated position is recognizing that you might be trying to solve the wrong kind of problem...
The real danger isn’t that machines will become intelligent—it’s that we’ll mistake impressive computation for understanding and surrender our judgment to those who control the servers.
The circus continues. The ground approaches. And some of us are paying attention to the actual distance.
"

Why I’m Betting Against the AGI Hype
An Engineer’s (Philosophical) Perspective.

My message for the new year around GenAI, don't believe the hype about AGI, it taking large number over jobs, replacing humans etcetera.
It also makes mistakes, hallucinates and you should check every outcome you re-use. At the same also don't believe the hype about GenAI being worthless, awful, bad, only being a stochastic parrot. That simply isn't true as anyone who regularly uses topmodels knows.

The whole thing is optimized for scams, deception and other criminal behavior:
- user interface that deceptively pretends it's a human you're talking to
- claims from companies highly exaggerate capabilities
companies and "experts" constantly hype "AGI" which they (funnily enough) do to both make investors greedier and spread fear and as a distraction because these algorithms can't actually do what they keep promising
- large-scale accounting and financial fraud (e.g. what Nvidia is doing with circular selling)
- biggest case of copyright infringement in history
Note: I think the underlying technology is really cool, and definitely has use cases and can be used for actually good things. But: some technology just has more downsides than upsides, and some should only be used by experts in controlled environments. Leaded gasoline, asbestos and chlorofluorocarbon are also all really cool technology.
In this case perhaps the techology itself doesn't do anything inherently bad, however the people making it are lying about what it can do, the people selling it are motivated purely by greed and the people using it (often forced to do so) are being deceived.

This website illustrates nicely how the US lost the competition–at least for now–in open(ish) LLM models: #AIResearch #AGI_hype
/via Wired