I Told You So - 8 September 2025

Not for nuthin, but AI (as currently implemented) is horseshit. You've read such here a few times before. But... now an AI veteran, is proposing (not for the first time) that the current approach can't work. I could have told you that. Wait... I did.

The current strategy of merely making A.I. bigger is deeply flawed — scientifically, economically and politically. Many things, from regulation to research strategy, must be rethought. One of the keys to this may be training and developing A.I. in ways inspired by the cognitive sciences.

IOW, what we learned in high school: GIGO.

Large language models, which power systems like GPT-5, are nothing more than souped-up statistical regurgitation machines, so they will continue to stumble into problems around truth, hallucinations and reasoning. Scaling would not bring us to the holy grail of A.G.I.

Ah. Another I told you so moment. GIGO. Again. A massive correlation matrix of shit. Way to go Bros.

The government has let A.I. companies lead a charmed life, with almost zero regulation. It now ought to enact legislation that addresses costs and harms unfairly offloaded onto the public — from misinformation to deepfakes, A.I. slop content, cybercrime, copyright infringement, mental health and energy usage.

And, now another I told you so moment: make 'em pay for their externalities. That'll cut into the profit. If there ever is any. (the link has a bunch more links in this sentence; way too lazy to C&P here.)

The cognitive sciences (including psychology, child development, philosophy of mind and linguistics) teach us that intelligence is about more than mere statistical mimicry and suggest three promising ideas for developing A.I. that is reliable enough to be trustworthy, with a much richer intelligence.

He goes on to describe those three ideas. You should read him up. But... I will crow that, taken together, they amount to a RDBMS of ground facts to constrain the correlation matrix. There. I said it. Again. His second idea is nearly explicit:

But as cognitive scientists like Steven Pinker, Elizabeth Spelke and me have emphasized, the human mind is born with some core knowledge of the world that sets us up to grasp more complex concepts.

QED Like any good maths proof.