malthaus

ycombinator (which, nota bene, also spawned sam) is nothing but distilled capitalism with a fig leaf to appeal to a subset of nerd culture and you all ate it up

i can understand them trying to prevent their business from becoming a commodity but i don't see that working out for them except for some short term buzz, but others will run with their ideas in domain specific applications

i need more compute? using a saas solution i need more storage? tiny external ssd or cloud storage

the 8gigs of ram also aren't holding me back, i think most people just cry based on principle and would be fine with it in a double-blind test

turning this into another overly verbose blog post doesn't really help my thinking on the matter (which, to be fair, is even less relevant than his musings)

the eu gets some publicity

and the public gets nothing but another bite out of their taxes

cryptocurrencies are a failure and net negative, and the longer we hold on to believing they aren't, the more damage we are doing

they eat ramen for 10 years, often even believing in someone else's dream without substantional equity to match the risk/reward profile. meanwhile, the plumber next door who started his own plumbing business is driving a ferrari on weekends.

sure, a full-on AGI, non-hallucinating AI would be great. but the current state is already a giant leap. there's so much untapped potential in the corporate world where whole departments, processes, etc can be decimated.

doing this and dealing with the socio-economic and political fall-out from those efficiency leaps can happen while research (along multiple pathways) goes on, and this will take 5-10 years at least.

everyone is looking at llm scores & strawberry gotchas while ignoring the trillions of market potential in replacing existing systems and (yes) people with the current capabilities. identifying the use cases, finetuning the models and (most importantly) actually rolling this out in existing organizations/processes/systems will be the challenge long before the base models' capabilities will be

it is worth working on those issues now and get the ball rolling, switching out your models for future more capable ones will be the easy part later on.

karpathy agrees with you, here he is hating on tokenizers while re-building them for 2h

but as someone who hates subscriptions, this would actually be a really nice subscription - currently i just buy the "all samples from [year - 1]" on black friday instead but i wouldn't mind a more automated / streamlined approach

as soon as one person starts reclininig, you can die on that hill all alone and i'm reclining faster than you can sigh in disappointment

i'm with the parent commenter here as i find your response very much telling on how you evaluate things and frankly, more idiotic than someone's life being changed, even by the "intended" interpretation of the book

e.g. would you say that everyone who liked / was influenced by reading "mein kampf" is a nazi?

the only way the valuation is put "in perspective" is if you estimate that we gonna have a race of heavy investments now to get to a point where the models are "good enough" and you no longer need to invest in training new ones, at which point you switch from running a immensely unprofitable business to a immensely profitable one.

the issue is that nobody knows when that will be the case (if ever) and it currently looks like whoever player is best equipped with capital to fund that race will have a winner takes all future ahead so you just place all your chips and hope for the best

because i'd consider myself above average in terms of intelligence and ambition but i still fall into the procrastination trap often. now you might say that this makes me in fact "stupid" per your defininition (or maybe arrogant as i overestimate myself) but i see this in other people as well.

i also would not say that being "productive" as in moving humanity ahead must be the KPI by which everyone is measured. you only have one life, you can spend it how you want, even if that is watching tiktoks 24/7.

if this house of cards collapses, it will be magnitutes bigger than any bubble (be it hype or fraud) we've seen so far.

what a waste of resources and focus, i really can't take vc players seriously any more (from what is left after the crypto hype run)