Simon Willison’s Weblog

There is something so vulnerable and frightening about doing your own thing, because it’s your fault if it doesn’t work. And then there’s this other kind of work, where you’re paid an extraordinary amount of money, you’re the hero before you walk in the door, you’re not even held that accountable, because you have a limited amount of time, and all you can do is make it better.

— Craig Mazin # 31st December 2023, 8:53 pm

Basically, we’re in the process of replacing our whole social back-end with ActivityPub. I think Flipboard is going to be the first mainstream consumer service that existed in a walled garden that switches over to ActivityPub.

— Mike McCue, CEO of Flipboard # 18th December 2023, 6:45 pm

Computer, display Fairhaven character, Michael Sullivan. [...]

Give him a more complicated personality. More outspoken. More confident. Not so reserved. And make him more curious about the world around him.

Good. Now... Increase the character’s height by three centimeters. Remove the facial hair. No, no, I don’t like that. Put them back. About two days’ growth. Better.

Oh, one more thing. Access his interpersonal subroutines, familial characters. Delete the wife.

— Captain Janeway, prompt engineering # 15th December 2023, 9:46 pm

And so the problem with saying “AI is useless,” “AI produces nonsense,” or any of the related lazy critique is that destroys all credibility with everyone whose lived experience of using the tools disproves the critique, harming the credibility of critiquing AI overall.

— Danilo Campos # 15th December 2023, 9:28 pm

gpt-4-turbo over the API produces (statistically significant) shorter completions when it “thinks” its December vs. when it thinks its May (as determined by the date in the system prompt).

I took the same exact prompt over the API (a code completion task asking to implement a machine learning task without libraries).

I created two system prompts, one that told the API it was May and another that it was December and then compared the distributions.

For the May system prompt, mean = 4298
For the December system prompt, mean = 4086

N = 477 completions in each sample from May and December

t-test p < 2.28e-07

— Rob Lynch # 11th December 2023, 7:45 pm

When I speak in front of groups and ask them to raise their hands if they used the free version of ChatGPT, almost every hand goes up. When I ask the same group how many use GPT-4, almost no one raises their hand. I increasingly think the decision of OpenAI to make the “bad” AI free is causing people to miss why AI seems like such a huge deal to a minority of people that use advanced systems and elicits a shrug from everyone else.

— Ethan Mollick # 10th December 2023, 8:17 pm

I always struggle a bit with I’m asked about the “hallucination problem” in LLMs. Because, in some sense, hallucination is all LLMs do. They are dream machines.

We direct their dreams with prompts. The prompts start the dream, and based on the LLM’s hazy recollection of its training documents, most of the time the result goes someplace useful.

It’s only when the dreams go into deemed factually incorrect territory that we label it a “hallucination”. It looks like a bug, but it’s just the LLM doing what it always does.

— Andrej Karpathy # 9th December 2023, 6:08 am

Create a culture that favors begging forgiveness (and reversing decisions quickly) rather than asking permission. Invest in infrastructure such as progressive / cancellable rollouts. Use asynchronous written docs to get people aligned (“comment in this doc by Friday if you disagree with the plan”) rather than meetings (“we’ll get approval at the next weekly review meeting”).

— Stay SaaSy # 8th December 2023, 6:21 pm

We like to assume that automation technology will maintain or increase wage levels for a few skilled supervisors. But in the long-term skilled automation supervisors also tend to earn less.

Here’s an example: In 1801 the Jacquard loom was invented, which automated silkweaving with punchcards. Around 1800, a manual weaver could earn 30 shillings/week. By the 1830s the same weaver would only earn around 5s/week. A Jacquard operator earned 15s/week, but he was also 12x more productive.

The Jacquard operator upskilled and became an automation supervisor, but their wage still dropped. For manual weavers the wages dropped even more. If we believe assistive AI will deliver unseen productivity gains, we can assume that wage erosion will also be unprecedented.

— Sebastian Majstorovic # 8th December 2023, 1:34 am

GPT and other large language models are aesthetic instruments rather than epistemological ones. Imagine a weird, unholy synthesizer whose buttons sample textual information, style, and semantics. Such a thing is compelling not because it offers answers in the form of text, but because it makes it possible to play text—all the text, almost—like an instrument.

— Ian Bogost # 5th December 2023, 8:29 pm

A calculator has a well-defined, well-scoped set of use cases, a well-defined, well-scoped user interface, and a set of well-understood and expected behaviors that occur in response to manipulations of that interface.

Large language models, when used to drive chatbots or similar interactive text-generation systems, have none of those qualities. They have an open-ended set of unspecified use cases.

— Anthony Bucci # 5th December 2023, 8:12 pm

So something everybody I think pretty much agrees on, including Sam Altman, including Yann LeCun, is LLMs aren’t going to make it. The current LLMs are not a path to ASI. They’re getting more and more expensive, they’re getting more and more slow, and the more we use them, the more we realize their limitations.

We’re also getting better at taking advantage of them, and they’re super cool and helpful, but they appear to be behaving as extremely flexible, fuzzy, compressed search engines, which when you have enough data that’s kind of compressed into the weights, turns out to be an amazingly powerful operation to have at your disposal.

[...] And the thing you can really see missing here is this planning piece, right? So if you try to get an LLM to solve fairly simple graph coloring problems or fairly simple stacking problems, things that require backtracking and trying things and stuff, unless it’s something pretty similar in its training, they just fail terribly.

[...] So that’s the theory about what something like Q* might be, or just in general, how do we get past this current constraint that we have?

— Jeremy Howard # 1st December 2023, 2:49 am

This is what I constantly tell my students: The hard part about doing a tech product for the most part isn’t the what beginners think makes tech hard — the hard part is wrangling systemic complexity in a good, sustainable and reliable way.

Many non-tech people e.g. look at programmers and think the hard part is knowing what this garble of weird text means. But this is the easy part. And if you are a person who would think it is hard, you probably don’t know about all the demons out there that will come to haunt you if you don’t build a foundation that helps you actively keeping them away.

— atoav # 30th November 2023, 9:18 pm

This is nonsensical. There is no way to understand the LLaMA models themselves as a recasting or adaptation of any of the plaintiffs’ books.

— U.S. District Judge Vince Chhabria # 26th November 2023, 4:13 am

To some degree, the whole point of the tech industry’s embrace of “ethics” and “safety” is about reassurance. Companies realize that the technologies they are selling can be disconcerting and disruptive; they want to reassure the public that they’re doing their best to protect consumers and society. At the end of the day, though, we now know there’s no reason to believe that those efforts will ever make a difference if the company’s “ethics” end up conflicting with its money. And when have those two things ever not conflicted?

— Lucas Ropek # 23rd November 2023, 8:41 pm

We have reached an agreement in principle for Sam Altman to return to OpenAI as CEO with a new initial board of Bret Taylor (Chair), Larry Summers, and Adam D’Angelo.

— @OpenAI # 22nd November 2023, 6:04 am

I remember that they [Ev and Biz at Twitter in 2008] very firmly believed spam was a concern, but, “we don’t think it’s ever going to be a real problem because you can choose who you follow.” And this was one of my first moments thinking, “Oh, you sweet summer child.” Because once you have a big enough user base, once you have enough people on a platform, once the likelihood of profit becomes high enough, you’re going to have spammers.

— Del Harvey # 22nd November 2023, 4:59 am

Sam Altman expelling Toner with the pretext of an inoffensive page in a paper no one read would have given him a temporary majority with which to appoint a replacement director, and then further replacement directors. These directors would, naturally, agree with Sam Altman, and he would have a full, perpetual board majority—the board, which is the only oversight on the OA CEO. Obviously, as an extremely experienced VC and CEO, he knew all this and how many votes he (thought he) had on the board, and the board members knew this as well—which is why they had been unable to agree on replacement board members all this time.

— Gwern # 22nd November 2023, 3:53 am

The way I think about the AI of the future is not as someone as smart as you or as smart as me, but as an automated organization that does science and engineering and development and manufacturing.

— Ilya Sutskever # 21st November 2023, 6:59 pm

And the investors wailed and gnashed their teeth but it’s true, that is what they agreed to, and they had no legal recourse. And OpenAI’s new CEO, and its nonprofit board, cut them a check for their capped return and said “bye” and went back to running OpenAI for the benefit of humanity. It turned out that a benign, carefully governed artificial superintelligence is really good for humanity, and OpenAI quickly solved all of humanity’s problems and ushered in an age of peace and abundance in which nobody wanted for anything or needed any Microsoft products. And capitalism came to an end.

— Matt Levine, in a hypothetical # 20th November 2023, 9:12 pm

The company pressed forward and launched ChatGPT on November 30. It was such a low-key event that many employees who weren’t directly involved, including those in safety functions, didn’t even realize it had happened. Some of those who were aware, according to one employee, had started a betting pool, wagering how many people might use the tool during its first week. The highest guess was 100,000 users. OpenAI’s president tweeted that the tool hit 1 million within the first five days. The phrase low-key research preview became an instant meme within OpenAI; employees turned it into laptop stickers.

— Inside the Chaos at OpenAI # 20th November 2023, 4:38 am

The EU AI Act now proposes to regulate “foundational models”, i.e. the engine behind some AI applications. We cannot regulate an engine devoid of usage. We don’t regulate the C language because one can use it to develop malware. Instead, we ban malware and strengthen network systems (we regulate usage). Foundational language models provide a higher level of abstraction than the C language for programming computer systems; nothing in their behaviour justifies a change in the regulatory framework.

— Arthur Mensch, Mistral AI # 16th November 2023, 11:29 am

I’ve resigned from my role leading the Audio team at Stability AI, because I don’t agree with the company’s opinion that training generative AI models on copyrighted works is ‘fair use’.

[...] I disagree because one of the factors affecting whether the act of copying is fair use, according to Congress, is “the effect of the use upon the potential market for or value of the copyrighted work”. Today’s generative AI models can clearly be used to create works that compete with the copyrighted works they are trained on. So I don’t see how using copyrighted works to train generative AI models of this nature can be considered fair use.

But setting aside the fair use argument for a moment — since ‘fair use’ wasn’t designed with generative AI in mind — training generative AI models in this way is, to me, wrong. Companies worth billions of dollars are, without permission, training generative AI models on creators’ works, which are then being used to create new content that in many cases can compete with the original works.

— Ed Newton-Rex # 15th November 2023, 9:31 pm

[On Meta’s Galactica LLM launch] We did this with a 8 person team which is an order of magnitude fewer people than other LLM teams at the time.

We were overstretched and lost situational awareness at launch by releasing demo of a *base model* without checks. We were aware of what potential criticisms would be, but we lost sight of the obvious in the workload we were under.

One of the considerations for a demo was we wanted to understand the distribution of scientific queries that people would use for LLMs (useful for instruction tuning and RLHF). Obviously this was a free goal we gave to journalists who instead queried it outside its domain. But yes we should have known better.

We had a “good faith” assumption that we’d share the base model, warts and all, with four disclaimers about hallucinations on the demo—so people could see what it could do (openness). Again, obviously this didn’t work.

— Ross Taylor # 15th November 2023, 1:15 am

Two things in AI may need regulation: reckless deployment of certain potentially harmful AI applications (same as any software really), and monopolistic behavior on the part of certain LLM providers. The technology itself doesn’t need regulation anymore than databases or transistors. [...] Putting size/compute caps on deep learning models is akin to putting size caps on databases or transistor count caps on electronics. It’s pointless and it won’t age well.

— François Chollet # 13th November 2023, 1:46 am

Did you ever wonder why the 21st century feels like we’re living in a bad cyberpunk novel from the 1980s?

It’s because these guys read those cyberpunk novels and mistook a dystopia for a road map. They’re rich enough to bend reality to reflect their desires. But we’re [sci-fi authors] not futurists, we’re entertainers! We like to spin yarns about the Torment Nexus because it’s a cool setting for a noir detective story, not because we think Mark Zuckerberg or Andreesen Horowitz should actually pump several billion dollars into creating it.

— Charles Stross # 11th November 2023, 1:09 am

One of my fav early Stripe rules was from incident response comms: do not publicly blame an upstream provider. We chose the provider, so own the results—and use any pain from that as extra motivation to invest in redundant services, go direct to the source, etc.

— Michael Schade # 5th November 2023, 10:53 pm

The thing nobody talks about with engineering management is this:

Every 3-4 months every person experiences some sort of personal crisis. A family member dies, they have a bad illness, they get into an argument with another person at work, etc. etc. Sadly, that is just life. Normally after a month or so things settle down and life goes on.

But when you are managing 6+ people it means there is *always* a crisis you are helping someone work through. You are always carrying a bit of emotional burden or worry around with you.

— Chris Albon # 27th October 2023, 6:18 am

If a LLM is like a database of millions of vector programs, then a prompt is like a search query in that database [...] this “program database” is continuous and interpolative — it’s not a discrete set of programs. This means that a slightly different prompt, like “Lyrically rephrase this text in the style of x” would still have pointed to a very similar location in program space, resulting in a program that would behave pretty closely but not quite identically. [...] Prompt engineering is the process of searching through program space to find the program that empirically seems to perform best on your target task.

— François Chollet # 25th October 2023, 11:26 pm

The real value in evolving as an engineer isn’t solely about amassing a heap of isolated skills but weaving them into an intricate web of abilities that’s greater than the sum of its parts.

— Addy Osmani # 24th October 2023, 6:09 am