20 Years Ago
In 2006 I basically finished the first draft of my Ph.D. thesis. It would drag on for a while because in 2006 this site was born and I spent over 100 hours per week on it. My Ph.D. was not activism; it dealt with MRI analysis for the most part - it did statistical analysis and synthesis - something which the media nowadays rushes to call "generative AI". So yes, one can say I did "generative AI" more than 2 decades ago - well before the hype. It was actually done for useful purposes. It wasn't slop and we didn't rely on whole datacentres for it. I'd use Fedora Core computer clusters overnight (in the Computer Science labs, over SSH).
The "generative AI" or "advanced" or "agentic" or "superintelligent" or whatever they'll call it tomorrow HYPE is ludicrous on many levels. Thankfully, we're seeing the beginning of the end of it. ChatGPT is down every month, Microsoft 'Open' 'AI' will soon run out of money (and no, it's not worth what the media claims; that media merely participates in financial fraud), there's no "better model" on the way, and a lot of the population has gotten truly fed up with slop.
Slop or "genAI" or whatever isn't new at all. A lot of what they call "AI" is just good old "CG" with more brute force (subsidised by companies that will only ever lose money).
Some time soon all this slop frenzy will become like yesterday's "blockchain" or "metaverse". █