Alex

Believer in AI & Humans beats AI alone
Interactive AI Tool Builder
⚡

[Alex @Contempl8t]

Decided to build a game to showcase why I think human-in-the-loop solutions need to improve for the full value of AI to be realized.

Introducing: https://t.co/h4iPUjW0EO 🚀

[Alex @Contempl8t]

AI is great for things you could have worked out yourself given enough time.

It's the rest where it's sketchy. We will need to constantly exceed ourselves in this new era.

[Alex @Contempl8t]

The takeaway isn't that intelligence is easy to replicate, it's that it's easy to fake.

[Alex @Contempl8t]

No one's or no AI is really really smart.

Even genius is only clear in hindsight because it's stumbled upon and it's never clear at the time how something will really end up vibing with the universe.

Master plans are interesting thought processes but things never go according to plan.

LLM intelligence doesn't solve any of this.

[Alex @Contempl8t]

@mal_shaik Does he present himself as the builder of the product or just "some guy" who uses the product.

[Alex @Contempl8t]

@timsoret In D2 mobs actually get stronger faster than you do. Like running Hell in hardcore is pretty insane.

But you know what, best shit ever IMO.

You actually had to strategize and hone your skills.

[Alex @Contempl8t]

@realwillreil Soon this will be normal and not a postable évent 😄

[dax @thdxr]

i said this to the team the other day, if we ever do a consumer product it should not be in the realm of productivity at all

[Alex @Contempl8t]

Frontier labs releasing cost effective models instead of major improvements and bigger context windows.
Does this signal something?

[Alex @Contempl8t]

@kunchenguid Or switch to cheaper model for late compact and switch back?

[Mehdi (e/λ) @BetterCallMedhi]

bon je considère que Laurent se trompre très largement avec cette take et je vais m’expliquer

déjà je pense que la + grande tragédie des 2 derniers siècles aura été de contraindre la conscience biologique à simuler le comportement d'un processeur, la bureaucratie, l'administration de surface, la saisie de données et la répétition mécanique etc etc.. (la globalité du travail tertiaire) n'ont jamais été conçues pour l'esprit humain

a mon sens ces métiers ont mutilé la dignité de l'être en ravalant des millions d'âmes au rang d'exécutants algorithmiques et nous avons qualifié de «progrès » un modèle de société dont l'unique produit était la déshumanisation du temps et l'asphyxie des relations réelles

d’ici 2030 je dirais même, l'intelligence artificielle ne va pas détruire le travail humain,elle va liquider le travail artificiel (tout le lot de BS jobs et m’a recommandation de lire David Graeber tient toujours) en ramenant le coût marginal de l'exécution mécanique et du jargon symbolique à 0 et en ce sens je dirais que l’IA va nous arracher notre masque de robots

désolé que je vais choquer mais je pense que tout ce que l'IA peut automatiser n'a jamais été digne de l'esprit humain et ne subsistera que ce qui résiste à la simulation: la présence physique, l'empathie brute, le courage, la créativité, la philosophie, l'art incarné et la beauté d'un lien social libéré de la transaction et cnest ce que je trouve absolument fabuleux

de manière plus profonde je crois que nous entrons dans une forme de renaissance existentialiste de notre histoire et contrairement à ceux qui parlent d’émettre une forme de conscience chez les robots, 2030 ne sera pas l'année où les machines seront enfin devenues humaines mais très certainement l'instant sacré où les êtres humains auront enfin le droit d'arrêter d'être des machines

[Yann LeCun @ylecun]

I said "auto-regressive LLMs, in and of themselves, will not lead human-level AI"
That statement is still totally true.

First, the reasoning abilities of current AI systems are based non-auto-regressive search (which is what I have always advocated for). But AFAICT, they do it in token space, which is limited and inefficient. I have claimed that human-like reasoning must be a search in continuous representation space. It looks like the industry is moving towards that.

Second, the self-improvement methods, as currently practiced, only work for domains where the quality of outputs can be scored without human intervention, such as mathematics, code, and scenarios that can be simulated accurately. Not anything else. Humans and animals learn new skills way more efficiently than current RL methods.

Third, the multimodal capabilities of current AI assistants generally use separately-trained encoders (that are not LLMs). This is also what I've been advocating. Except that I think the best way to do this is with JEPA trained with self-supervised learning. The research community is clearly moving towards that (3000 papers on JEPA in just 4 years).

Fourth, if LLMs were a path to human-level AI, we would have domestic robots and Level-4 or Level-5 self-driving cars for consumers by now. And we don't. We certainly don't have cars that can learn to drive in 20 hours or practice like any teenager. We're still missing something pretty huge to claim human-level intelligence (let alone superhuman).

Sure, we now have computer systems that are impressive, very useful, and whose performance is superhuman in an increasing number of domains (coding being one of them).
But that's true of the entire history of progress in computer technology.

Lastly, there is a basic confusion about what intelligence actually is.
It is not the mere accumulation and regurgitation of existing declarative knowledge (which is essentially what LLMs do).
As Jean Piaget famously said, "intelligence is not what you know, it is what you do when you don't know."
It is your ability to solve new problem without any prior training, to act in previously-unknown scenarios, and to adapt very quickly to new situations with minimal training.

We're still far from that.

[Alex @Contempl8t]

Sometimes when you are too close to a topic and the details, you lose sight of the bigger picture and what's actually important.

[roon @tszzl]

@souljagoyteller this sounds completely incoherent to me

[Alex @Contempl8t]

@Satyaki_R You need to be in an industry where AI constantly disappoints.

[Alex @Contempl8t]

@effectfully Not sure why this isn't said more often.

This type of stuff happens all the time and people keep on talking about the amazing intelligence when it's so clear there's something very fundamental missing.

[Alex @Contempl8t]

@themgmtconsult Yeah... The early wake up is definitely a sure way lol

[FleetingBits @fleetingbits]

some thoughts on jev

1) i tend to think that twitter hype is not relevant for ai product releases; especially when someone claims to be doing something novel with the architecture

2) but, i think that jev is interesting in principle; it is a low cost, fast classification service, and it claims to be about equivalent to gpt-5.6-terra in performance

3) something interesting here is that this is what most ai services looked like in 2022; cohere sold a classification api, an embedding api, a reranker

4) the problem was that, in addition to not being as effective as gpt-3 and then gpt-4, these services were much harder to use as a developer

5) you had to decide which apis to use, then write the code to glue them together; so, you got less performance for more work

6) but now, we can imagine coding agents making it much easier to select and integrate a cheaper service with specialized characteristics

7) evals are still a bottleneck, though; you need some way to automatically compare performance between services using their apis

8) and, enterprise companies often have a hard time building evals that reflect their actual use cases; it's not a core expertise

9) but, the better coding agents get at integrating and switching between services, the more valuable those evals become;

10) this means companies have more reason to figure out the evals problem;

11) maybe that means getting rights to use customer data for testing, or getting better at purchasing synthetic test data, or something else

12) so, i think there is an interesting world where coding agents create more market for specialized api services, even specialized model services

13) which can compete either on cost or on capabilities; with the winner determined by cost basis or specialized data, etc...

14) note, it might be unclear but if specialized models services becomes an important direction, i expect the frontier labs to have an edge on cost basis

15) and, i think that, at least for now, these specialized model services probably do not provide enough of a cost difference to be relevant

[Alex @Contempl8t]

@conksresearch "will delete if wrong"

You can recognize a good trader by their love for options 😂

[Alex @Contempl8t]

@realwillreil Very useful if you make a profit 😂

[Alex @Contempl8t]

@ThePrimeagen I remember when you used to think decision speed couldn't follow LLM speed. I guess you really did make it to 100x 😉