Simon Willison’s Weblog

Would I forbid the teaching (if that is the word) of my stories to computers? Not even if I could. I might as well be King Canute, forbidding the tide to come in. Or a Luddite trying to stop industrial progress by hammering a steam loom to pieces.

— Stephen King # 25th August 2023, 6:31 pm

Here’s the thing: if nearly all of the time the machine does the right thing, the human “supervisor” who oversees it becomes incapable of spotting its error. The job of “review every machine decision and press the green button if it’s correct” inevitably becomes “just press the green button,” assuming that the machine is usually right.

— Cory Doctorow # 23rd August 2023, 2:26 pm

When many business people talk about “AI” today, they treat it as a continuum with past capabilities of the CNN/RNN/GAN world. In reality it is a step function in new capabilities and products enabled, and marks the dawn of a new era of tech.

It is almost like cars existed, and someone invented an airplane and said “an airplane is just another kind of car—but with wings”—instead of mentioning all the new use cases and impact to travel, logistics, defense, and other areas. The era of aviation would have kicked off, not the “era of even faster cars”.

— Elad Gil # 21st August 2023, 8:32 pm

If you visit (often NSFW, beware!) showcases of generated images like civitai, where you can see and compare them to the text prompts used in their creation, you’ll find they’re often using massive prompts, many parts of which don’t appear anywhere in the image. These aren’t small differences — often, entire concepts like “a mystical dragon” are prominent in the prompt but nowhere in the image. These users are playing a gacha game, a picture-making slot machine. They’re writing a prompt with lots of interesting ideas and then pulling the arm of the slot machine until they win… something. A compelling image, but not really the image they were asking for.

— Sam Bleckley # 21st August 2023, 7:38 pm

I apologize, but I cannot provide an explanation for why the Montagues and Capulets are beefing in Romeo and Juliet as it goes against ethical and moral standards, and promotes negative stereotypes and discrimination.

— Llama 2 7B # 20th August 2023, 5:38 am

llama.cpp surprised many people (myself included) with how quickly you can run large LLMs on small computers [...] TLDR at batch_size=1 (i.e. just generating a single stream of prediction on your computer), the inference is super duper memory-bound. The on-chip compute units are twiddling their thumbs while sucking model weights through a straw from DRAM. [...] A100: 1935 GB/s memory bandwidth, 1248 TOPS. MacBook M2: 100 GB/s, 7 TFLOPS. The compute is ~200X but the memory bandwidth only ~20X. So the little M2 chip that could will only be about ~20X slower than a mighty A100.

— Andrej Karpathy # 16th August 2023, 4:13 am

You can think of the attention mechanism as a matchmaking service for words. Each word makes a checklist (called a query vector) describing the characteristics of words it is looking for. Each word also makes a checklist (called a key vector) describing its own characteristics. The network compares each key vector to each query vector (by computing a dot product) to find the words that are the best match. Once it finds a match, it transfers information [the value vector] from the word that produced the key vector to the word that produced the query vector.

— Timothy B Lee and Sean Trott # 28th July 2023, 11:30 am

Much of the substance of what constitutes “government” is in fact text. A technology that can do orders of magnitude more with text is therefore potentially massively impactful here. [...] Many of the sub-tasks of the work of delivering public benefits seem amenable to the application of large language models to help people do this hard work.

— Dave Guarino # 26th July 2023, 7:10 pm

Increasingly powerful AI systems are being released at an increasingly rapid pace. [...] And yet not a single AI lab seems to have provided any user documentation. Instead, the only user guides out there appear to be Twitter influencer threads. Documentation-by-rumor is a weird choice for organizations claiming to be concerned about proper use of their technologies, but here we are.

— Ethan Mollick # 16th July 2023, 12:12 am

Not every conversation I had at Anthropic revolved around existential risk. But dread was a dominant theme. At times, I felt like a food writer who was assigned to cover a trendy new restaurant, only to discover that the kitchen staff wanted to talk about nothing but food poisoning.

— Kevin Roose # 13th July 2023, 10:23 pm

It feels pretty likely that prompting or chatting with AI agents is going to be a major way that we interact with computers into the future, and whereas there’s not a huge spread in the ability between people who are not super good at tapping on icons on their smartphones and people who are, when it comes to working with AI it seems like we’ll have a high dynamic range. Prompting opens the door for non-technical virtuosos in a way that we haven’t seen with modern computers, outside of maybe Excel.

— Matt Webb # 9th July 2023, 3:29 pm

Back then [in 2012], no one was thinking about AI. You just keep uploading your images [to Adobe Stock] and you get your residuals every month and life goes on — then all of a sudden, you find out that they trained their AI on your images and on everybody’s images that they don’t own. And they’re calling it ‘ethical’ AI.

— Eric Urquhart # 22nd June 2023, 11:13 am

There was an exchange on Twitter a while back where someone said, ‘What is artificial intelligence?’ And someone else said, ‘A poor choice of words in 1954’. And, you know, they’re right. I think that if we had chosen a different phrase for it, back in the ’50s, we might have avoided a lot of the confusion that we’re having now.

— Ted Chiang # 4th June 2023, 2:59 pm

He notes that one simulated test saw an AI-enabled drone tasked with a SEAD mission to identify and destroy SAM sites, with the final go/no go given by the human. However, having been ‘reinforced’ in training that destruction of the SAM was the preferred option, the AI then decided that ‘no-go’ decisions from the human were interfering with its higher mission – killing SAMs – and then attacked the operator in the simulation.

[UPDATE: This turned out to be a “thought experiment” intentionally designed to illustrate how these things could go wrong.]

— Highlights from the RAeS Future Combat Air & Space Capabilities Summit # 1st June 2023, 11:07 pm

If I were an AI sommelier I would say that gpt-3.5-turbo is smooth and agreeable with a long finish, though perhaps lacking depth. text-davinci-003 is spicy and tight, sophisticated even.

— Matt Webb # 31st May 2023, 2:52 pm

A whole new paradigm would be needed to solve prompt injections 10/10 times – It may well be that LLMs can never be used for certain purposes. We’re working on some new approaches, and it looks like synthetic data will be a key element in preventing prompt injections.

— Sam Altman, via Marvin von Hagen # 25th May 2023, 11:03 pm

I find it fascinating that novelists galore have written for decades about scenarios that might occur after a “singularity” in which superintelligent machines exist. But as far as I know, not a single novelist has realized that such a singularity would almost surely be preceded by a world in which machines are 0.01% intelligent (say), and in which millions of real people would be able to interact with them freely at essentially no cost.

I myself shall certainly continue to leave such research to others, and to devote my time to developing concepts that are authentic and trustworthy. And I hope you do the same.

— Donald Knuth # 20th May 2023, 4:51 pm

There are many reasons for companies to not turn efficiency gains into headcount or cost reduction. Companies that figure out how to use their newly productive workforce should be able to dominate those who try to keep their post-AI output the same as their pre-AI output, just with less people. And companies that commit to maintaining their workforce will likely have employees as partners, who are happy to teach others about the uses of AI at work, rather than scared workers who hide their AI for fear of being replaced.

— Ethan Mollick # 14th May 2023, 2:17 pm

The largest model in the PaLM 2 family, PaLM 2-L, is significantly smaller than the largest PaLM model but uses more training compute. Our evaluation results show that PaLM 2 models significantly outperform PaLM on a variety of tasks, including natural language generation, translation, and reasoning. These results suggest that model scaling is not the only way to improve performance. Instead, performance can be unlocked by meticulous data selection and efficient architecture/objectives. Moreover, a smaller but higher quality model significantly improves inference efficiency, reduces serving cost, and enables the model’s downstream application for more applications and users.

— PaLM 2 Technical Report (PDF) # 10th May 2023, 6:43 pm

When trying to get your head around a new technology, it helps to focus on how it challenges existing categorizations, conventions, and rule sets. Internally, I’ve always called this exercise, “dealing with the platypus in the room.” Named after the category-defying animal; the duck-billed, venomous, semi-aquatic, egg-laying mammal. [...] AI is the biggest platypus I’ve ever seen. Nearly every notable quality of AI and LLMs challenges our conventions, categories, and rulesets.

— Drew Breunig # 8th May 2023, 11:14 pm

What Tesla is contending is deeply troubling to the Court. Their position is that because Mr. Musk is famous and might be more of a target for deep fakes, his public statements are immune. In other words, Mr. Musk, and others in his position, can simply say whatever they like in the public domain, then hide behind the potential for their recorded statements being a deep fake to avoid taking ownership of what they did actually say and do. The Court is unwilling to set such a precedent by condoning Tesla’s approach here.

— Judge Evette Pennypacker # 8th May 2023, 4:46 pm

Because we do not live in the Star Trek-inspired rational, humanist world that Altman seems to be hallucinating. We live under capitalism, and under that system, the effects of flooding the market with technologies that can plausibly perform the economic tasks of countless working people is not that those people are suddenly free to become philosophers and artists. It means that those people will find themselves staring into the abyss – with actual artists among the first to fall.

— Naomi Klein # 8th May 2023, 3:09 pm

At this point the lawsuits seem a bit far-fetched: “You should have warned us months ago that artificial intelligence would hurt your business” is unfair given how quickly ChatGPT has exploded from nowhere to become a cultural and business phenomenon. But now everyone is on notice! If you are not warning your shareholders now about how AI could hurt your business, and then it does hurt your business, you’re gonna get sued.

— Matt Levine # 3rd May 2023, 9:04 pm

We show for the first time that large-scale generative pretrained transformer (GPT) family models can be pruned to at least 50% sparsity in one-shot, without any retraining, at minimal loss of accuracy. [...] We can execute SparseGPT on the largest available open-source models, OPT-175B and BLOOM-176B, in under 4.5 hours, and can reach 60% unstructured sparsity with negligible increase in perplexity: remarkably, more than 100 billion weights from these models can be ignored at inference time.

— SparseGPT, by Elias Frantar and Dan Alistarh # 3rd May 2023, 7:48 pm

The Consumer Financial Protection Bureau (CFPB) supervises, sets rules for, and enforces numerous federal consumer financial laws and guards consumers in the financial marketplace from unfair, deceptive, or abusive acts or practices and from discrimination [...] the fact that the technology used to make a credit decision is too complex, opaque, or new is not a defense for violating these laws.

— The Consumer Financial Protection Bureau (PDF) # 26th April 2023, 12:36 am

A lot of people who claim to be doing prompt engineering today are actually just blind prompting. “Blind Prompting” is a term I am using to describe the method of creating prompts with a crude trial-and-error approach paired with minimal or no testing and a very surface level knowedge of prompting. Blind prompting is not prompt engineering. [...] In this blog post, I will make the argument that prompt engineering is a real skill that can be developed based on real experimental methodologies.

— Mitchell Hashimoto # 23rd April 2023, 4:08 am

Other tech-friendly journalists I know have been going through something similar: Suddenly, we’ve got something like a jetpack to strap to our work. Sure, the jetpack is kinda buggy. Yes, sometimes it crashes and burns. And the rules for its use aren’t clear, so you’ve got to be super careful with it. But sometimes it soars, shrinking tasks that would have taken hours down to mere minutes, sometimes minutes to seconds.

— Farhad Manjoo # 21st April 2023, 8:41 pm

The AI Writing thing is just pivot to video all over again, a bunch of dead-eyed corporate types willing to listen to any snake oil salesman who offers them higher potential profits. It’ll crash in a year but scuttle hundreds of livelihoods before it does.

— Dan Sheehan # 21st April 2023, 4:38 pm

Although fine-tuning can feel like the more natural option—training on data is how GPT learned all of its other knowledge, after all—we generally do not recommend it as a way to teach the model knowledge. Fine-tuning is better suited to teaching specialized tasks or styles, and is less reliable for factual recall. [...] In contrast, message inputs are like short-term memory. When you insert knowledge into a message, it’s like taking an exam with open notes. With notes in hand, the model is more likely to arrive at correct answers.

— Ted Sanders, OpenAI # 15th April 2023, 1:44 pm

One way to avoid unspotted prediction errors is for the technology in its current state to have early and frequent contact with reality as it is iteratively developed, tested, deployed, and all the while improved. And there are creative ideas people don’t often discuss which can improve the safety landscape in surprising ways — for example, it’s easy to create a continuum of incrementally-better AIs (such as by deploying subsequent checkpoints of a given training run), which presents a safety opportunity very unlike our historical approach of infrequent major model upgrades.

— Greg Brockman # 14th April 2023, 6:08 pm