Seriously, you read the headline, you say "Oh shit", & then you remember, this is the 20th to 30th time in the last year that OpenAI or Antropic - and now Nvidia - has made some overblown and/or alarming post about the state of GenAI. Maybe claiming we are at AGI - which we totally aren't.
LLMs. aka Autocomplete On Steroids, are still, from the last I heard, producing at a minimum 9% error rates.
Agents build on that to get to a 15% MINIMUM error rate.
YOU CANNOT BUILD RELIABLE SYSTEMS WITH ERROR RATES THAT HIGH.
Yes, the stuff that GenAI produces can be impressive, but, still, it is by definition BULLSHIT: it looks like it might be credible, but, maybe it is completely & utterly false - completely made up out of statistics, with 0 real knowledge of anything. Anyone producing AI slop output without serious human fact-checking should be completely liable for criminal charges.
LLMs are sucn an interesting technology. It's incredible that 1 algorithm, the Transformer, discovered by the Google Translate team, could create such an upheaval.
It's so tragic that psychopathic TechBros decided that this was their ticket to generational wealth.
LLMs are a normal technology. They are a great tool for experts who can readily check their output. For everyone else, they are Acme Slop Generators (going into my table of names).
Wow, Small Language Models running on PCs or phones are 98.6% as accurate as LLMs running on $B of hardware??? Oops.
From Joachim Klement on SubStack:
If this is true, the hyperscalers are toastAs you can see, in every domain, the best SLM is able to find the same or better answers than an LLM in 90% or more of the cases, with an average across all domains of 98.6%.
From the Bard of the Revolution, Cory Doctorow:
The actual epistemic crisis:AI's assault on reality is an opportunistic infection.
AI is alarming for many reasons: it's a dangerous financial bubble, an environmental catastrophe, and a tool for eroding wages and labor power. But in addition to all that, AI is an epistemic disaster.
We've had photoshopped images, voice impersonators and visual effects for years, of course, but with deepfakes, we've democratized access to reality-bending images, sounds and videos that appear real but are not. It's harder than ever to know what's true. Politicians and celebrities and activists show up in our feeds, declaring their fealty to this cause or product, or their fury at some turn in the world's events. Battlefields mound high with bodies and influencers marvel at impossible, sumptuous meals. It all seems plausible, and some of it is real, but not all of it, and because we know some of it is fake, we can't be sure if any of it isn't.
Again from Doctorow, somewhat more abstract. When you were born determines what doors are open for you?!?!?
Born on technology's third base:Material forces shape life-chances.
...
By the same token, continuing to give trillions to AI companies because they experienced early success with theory-free inference at scale tells us nothing about how to solve the vast range of problems that theory-free inference at scale sucks at. Doubling down on AI to overcome its increasingly obvious limitations is like doubling down on building post-war suburbs to fix today's housing market.
It's possible to achieve impressive feats because you're smart and hard working and also because you were in the right place at the right time. Historical contingency produced the AI bubble, and it is producing the conditions for that bubble to pop.
OK, that clears the links from my Lexington PC, from just before we went on vacation. I will next switch over to the links on my MacBook, from while we were gone.
Here's the home page/directory for my posts on Bullshit. This is post #143.
