TriProf

I'm a busy mom, professor, and ultimate age-grouper (as in, not very fast) who runs and bikes, mostly off-road. I don't train enough, but I do what I can.

[Jonathan Marks @marksjo1]

Departments often have “retreats.”Weak. This year. our department is having an “attack.” We shot our way into the conference room of the department of economics (they have a water cooler!) and have repelled efforts to dislodge us. This has energized and unified our department

[TriProf @triprof]

I can't reply to this account, so am using the quote feature instead: https://t.co/p4AKu8kceP

[Gary Marcus @GaryMarcus]

So many people are confused about the relation between human cognitive errors and LLM hallucinations that I wrote this short explainer:

Humans say things that aren't true for many different reasons
• Sometimes they lie
• Sometimes they misremember things
• Sometimes they fail to think through what they are saying
• Sometimes they are on drugs
• Sometimes they suffer from mental disorders
etc

LLMs errors result from 𝙖 𝙙𝙞𝙛𝙛𝙚𝙧𝙚𝙣𝙩 𝙪𝙣𝙙𝙚𝙧𝙡𝙮𝙞𝙣𝙜 𝙥𝙧𝙤𝙘𝙚𝙨𝙨. They don't have (e.g.,) intentions, egos, or financial interests, so they don't lie. They don't take drugs. They don't have emotional states.

Instead, LLM "hallucinations" arise, regularly, because (a) they literally don't know the difference between truth and falsehood, (b) they don't have reliably reasoning processes to guarantee that their inferences are correct and (c) they are incapable of fact-checking their own work. Instead, everything that LLMs say -- true or false -- comes from the same process of statistically reconstructing what words are likely in some context. They NEVER fact-check what they say. Some of it is true; some is false. But even with perfect data, the stochastic reconstructive process would still produce some errors. The very process that LLMs use to generalize also creates hallucinations. (In my 2001 book I explain what a different generalization process might look like.)

§

Importantly, the goal of AGI is not to recreate humans; we don't want AGI to lie or suffer from psychiatric disorders, for example. Rather, the goal of AGI should be to build machines that can reliably reason and plan about a wide swathe of the world. The fact that humans sometimes make errors, sometimes deliberately, sometimes accidentally, in no way takes away from -- or repairs -- the limitations of the current approach.

The field of AI will eventually do better, but probably with an AI that is structured differently, in which facts are first-class citizens, rather than something you hope you might get for free with enough data.

TL;DR: Don't console yourself with making something that superficially looks like human errors, if you aspire to AGI.

[TriProf @triprof]

I'm not the least bit German, but have never had a problem with stererotypical 'Germanness'. (Took the German or Autistic diagnostic. Result: Both. The Wittgenstein Result.) https://t.co/Rqr1kgo3JY

[TriProf @triprof]

@Paw_Vibe Bea

[TriProf @triprof]

Husband just poured me this monster glass of wine to deal with the game. #GoBills

[Andrea Montanari @Andrea__M]

The risk of AI for education is not students cheating in exams, it is people in general cheating themselves into believing they understand things they don’t.

[Sam Cooper @scoopercooper]

Don’t let it stand. Maybe Joe Tay will have a story to tell on The Bureau Podcast. Retweet if you want the real story of last election.

[TriProf @triprof]

It's kind of amazing to be able to say I've now eaten dinner with two Nobel Prize winners (in different fields), & had drinks with a third. (Also ridden in the car of a fourth, but it was not their car by that point in time.) Not bad for a kid from small town BC.

[TriProf @triprof]

@TristinHopper https://t.co/uuzSK3KRoF

[TriProf @triprof]

Rafflebox: Help support the DEVO/Escape Velocity Cycling Club and have a chance to win some real prizes. https://t.co/PkrjSMWsph

[TriProf @triprof]

Help support youth cycling in Vancouver and get a chance to win some cash! Devo's having their first 50/50 draw! https://t.co/PkrjSMWsph

[TriProf @triprof]

@benryanwriter Seems like something @Snack_Memories might be able to help with

[TriProf @triprof]

Splurged on these today.

[Alexandra Klasinski @alexandrak]

Happy Barkley Marathons to everyone who celebrates! #BM100

[TriProf @triprof]

@rcousine Oh no! So sorry for you and Pinter’s other human.

[TriProf @triprof]

@stuartlosaltos I should also specify that I make it with young turnip leaves. Mature ones I use like collards.

[TriProf @triprof]

@stuartlosaltos Hazelnuts. And I think I made it without cheese. I used fish sauce for the umami. (it's my go to secret ingredient.)

[Rochester, NY Buffalo Bills Backers @BillsBackersROC]

Spotted at Herrema's Food Market in Irondequoit. Go Bills!

#BillsMafia #GoBills #ROC #SuperBowl