I think it originated in this piece by Ted Chiang a couple years ago.

How is the hospital food?

Yeah I agree. Small models is the way. You can also use LoRa/QLoRa adapters to “fine tune” the same big model for specific tasks and swap the use case in realtime. This is what apple do with apple intelligence. You can outperform a big general LLM with an SLM if you have a nice specific use case and some data (which you can synthesise in come cases)

Unlike the dotcom bubble, Another big aspect of it is the unit cost to run the models.

Traditional web applications scale really well. The incremental cost of adding a new user to your app is basically nothing. Fractions of a cent. With LLMs, scaling is linear. Each machine can only handle a few hundred users and they’re expensive to run:

Big beefy GPUs are required for inference as well as training and they require a large amount of VRAM. Your typical home gaming GPU might have 16gb vram, 32 if you go high end and spend $2500 on it (just the GPU, not the whole pc). Frontier models need like 128gb VRAM to run and GPUs manufactured for data centre use cost a lot more. A state of the art Nvidia h200 costs $32k. The servers that can host one of these big frontier models cost, at best, $20 an hour to run and can only handle a handful of user requests so you need to scale linearly as your subscriber count increases. If you’re charging $20 a month for access to your model, you are burning a user’s monthly subscription every hour for each of these monster servers you have turned on. That’s generous and assumes they’re not paying the “on-demand” price of $60/hr.

Sam Altman famously said OpenAI are losing money on their $200/mo subscriptions.

If/when there is a market correction, a huge factor of the amount of continued interest (like with the internet after dotcom) is whether the quality of output from these models reflects the true, unsubsidized price of running them. I do think local models powered by things like llamacpp and ollama and which can run on high end gaming rigs and macbooks might be a possible direction for these models. Currently though you can’t get the same quality as state-of-the-art models from these small, local LLMs.

This is nice to see. I wish they’d actually do something about our water companies too.

We all have to feel sorry for the drones?

Relative privation is when someone dismisses or minimizes a problem simply because worse problems exist: “You can’t complain about X when Y exists.”

I’m talking about the practical reality that you must prioritize among legitimate problems. If you’re marooned at sea in a sinking ship you need to repair the hull before you try to fix the engines in order to get home.

It’s perfectly valid to say “I can’t focus on everything so I will focus on the things that provide the biggest and most tangible improvement to my situation first”. It’s fallacious to say “Because worse things exist, AGI concerns doesn’t matter.”

Here’s how I see it: we live in an attention economy where every initiative with a slew of celebrities attached to it is competing for eyeballs and buy in. It adds to information fatigue and analysis paralysis . In a very real sense if we are debating AGI we are not debating the other stuff. There are only so many hours in a day.

If you take the position that AGI is basically not possible or at least many decades away (I have a background in NLP/AI/LLMs and I take this view - not that it’s relevant in the broader context of my comment) then it makes sense to tell people to focus on solving more pressing issues e.g. nascent fascism, climate collapse, late stage capitalism etc.

I remember a friend telling me their English teacher took them on a field trip to see the Poet Simon Armitage and asked him what the poem “the hitcher” represents and he just said “it’s just about some bloke” - or something to that effect. Anyway it made me chuckle to imagine the dismay of the English teachers in the room

If we get a breakthrough moment with quantum, the machines will not be evenly distributed to start with. They will be too expensive to build, power and cool unless you’re a fortune 500 exactly like LLMs right now (aside from small models like llama that can run on consumer hardware). At the moment quantum computers rely on superconductors that have to be cooled near absolute zero which is… somewhat expensive to achieve.

Unlike LLMs (oh no I can’t talk to waifu without cell coverage waah) Not being able to run quantum algorithms on your phone in this scenario would be bad. It either means your personal comms are, for all intents and purposes decryptable by those who control the quantum machines or that you’ll have to pay rent to the people who control quantum machines to have them encrypt and decrypt stuff for you. Of course you’ll have to trust them too. Also, given governments thirst for spying on our encrypted comms, it’s possible that quantum machines are heavily regulated allowing “the good guys” a back door into our chats without giving “the baddies” a way to encrypt their comms

Everyone knows Franz Ferdinand, the indie rock band from Glasgow. Their 2004 hit “take me out” is definitely historical /s

Presumably it’s selling snake oil and convincing people to trust them?

In the United Kingdom yes because of our authoritarian Online Safety Act that came into power earlier this year. If I join a discord channel marked as nsfw I get a prompt for id which I bypass with a VPN in another country.

Same take my friend. I agree - Tim’s personal politics are kinda irrelevant in this context. Best for apple=compliance with whoever is in charge so they get to keep their money printer. Corpos gonna corpo

You’re giving apple too much credit. Tim Apple literally gifted trump a golden statue. The press are reporting that this is directly in response to a request from the DOJ. This isn’t apple having a moral epiphany and thinking they’re doing ice targets a favour. Corporations don’t do that sort of thing unless it will make them $$$

My Gmail account is old enough to buy drinks in a bar in the US (21)

Love him in any roles where he uses his regular British accent. Let’s not make him do Baltimore accent ever again. In fact, probably best we forget that happened l tbh

Sure they’re here

That’s what they’re saying. They’re dependent on cloudflare who offer a DNS service that routes traffic to one of their static ips, down a tunnel initiated by the server without an IP address.

At the lower end of the budget you could consider libreboot - it’s a one person band who ships refurbished Lenovo thinkpads with Linux pre-installed