Quotes

A man whose desire is to be something separate from himself, to be a member of Parliament, or a successful grocer, or a prominent solicitor, or a judge, or something equally tedious, invariably succeeds in being what he wants to be. That is his punishment. Those who want a mask have to wear it.

I think that a lot of resistance to AI coding tools comes from the same place: fear of losing something that has defined you for so long. People are reacting against overblown hype, and there is overblown hype. I get that, but I also think there’s something deeper going on here. When you’ve worked hard to build your skills, when coding is part of your identity and where you get your worth, the idea of a tool that might replace some of that is very threatening.

If I tweeted a throwaway comment in appreciation for McDonald’s apple pies and some other randos on Twitter happened to also tweet similar thoughts over the last few months, it doesn’t mean by extrapolation that ‘Millennials Can’t Get Enough Of McDonald’s Apple Pies’.  The Twitter search box is not a polling agency and Twitter doesn’t include everybody’s thoughts on everything. Just some people’s thoughts on some things.

IX (1914)—In complete helplessness wrote barely 2 pages. I have retreated considerably today, even though I had slept well. But I know that I must not yield, if I want to rise above the lowest woes of my writing, which is already held down by the rest of my way of life, into the greater freedom that might be waiting for me. The old dullness has not yet completely left me I realize and the coldness of my heart might never leave me. The fact that I recoil from no humiliation can just as well mean hopelessness as give hope.

Like DOCTYPE switching did in 2000, version targeting negates the vendor argument that existing behaviors can't be changed for fear of breaking web sites. If IE8 botches its implementation of some CSS property or DOM method, the mistake can be fixed in IE9 without breaking sites developed in the IE8 era. This actually makes browser vendors more susceptible to pressure to fix their bugs, and less fearful of doing so.

So much of knowledge/intelligence involves translating ideas between fields (domains). Those domains are walls the keep ideas siloed. But LLMs can help break those walls down and encourage humans to do more interdisciplinary thinking, which may lead to faster discoveries. And note that I am implying that humans will make the breakthroughs, using LLMs as translation tools when appropriate, to help make connections. LLMs are strongest as translators of information that you provide. BYOD: Bring your own data!

OpenAI's new o3 system - trained on the ARC-AGI-1 Public Training set - has scored a breakthrough 75.7% on the Semi-Private Evaluation set at our stated public leaderboard $10k compute limit. A high-compute (172x) o3 configuration scored 87.5%. This is a surprising and important step-function increase in AI capabilities, showing novel task adaptation ability never seen before in the GPT-family models. For context, ARC-AGI-1 took 4 years to go from 0% with GPT-3 in 2020 to 5% in 2024 with GPT-4o. All intuition about AI capabilities will need to get updated for o3.

I used to have this singular focus on students writing code that they submit, and then I run test cases on the code to determine what their grade is. This is such a narrow view of what it means to be a software engineer, and I just felt that with generative AI, I’ve managed to overcome that restrictive view. It’s an opportunity for me to assess their learning process of the whole software development [life cycle]—not just code. And I feel like my courses have opened up more and they’re much broader than they used to be. I can make students work on larger and more advanced projects.

Indy: "So you said you’re a good mimic?" Nicholas Gurewitch: [interrupting] "Not a good one: a patient one. I guess patience is very mistakable for talent these days...I think if you do anything patiently people mistake it for being genius when in reality it’s just the result of an unfathomable amount of patience, in my case."

Basically any resource on a difficult subject—a colleague, Google, a published paper—will be wrong or incomplete in various ways. Usefulness isn’t only a matter of correctness. For example, suppose a colleague has a question she thinks I might know the answer to. Good news: I have some intuition and say something. Then we realize it doesn’t quite make sense, and go back and forth until we converge on something correct. Such a conversation is full of BS but crucially we can interrogate it and get something useful out of it in the end. Moreover this kind of back and forth allows us to get to the key point in a way that might be difficult when reading a difficult ~50-page paper. To be clear o3-mini-high is orders of magnitude less useful for this sort of thing than talking to an expert colleague. But still useful along similar dimensions (and with a much broader knowledge base).

On two occasions I have been asked,—'Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?' In one case a member of the Upper, and in the other a member of the Lower, House put this question. I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.

No matter what great leaps forward the Internet Explorer team make from now on, the majority of developers won’t use them and the majority of users won’t see them. By doing this the Internet Explorer team may have created their own backwater, shot themselves in the foot and left themselves for dead.

Glory be to God for dappled things— For skies of couple-colour as a brinded cow; For rose-moles all in stipple upon trout that swim; Fresh-firecoal chestnut-falls; finches’ wings; Landscape plotted and pieced—fold, fallow, and plough; And áll trádes, their gear and tackle and trim.All things counter, original, spare, strange; Whatever is fickle, freckled (who knows how?) With swift, slow; sweet, sour; adazzle, dim; He fathers-forth whose beauty is past change: Praise him.

In too many organizations, deploy code is a technical backwater, an accumulation of crufty scripts and glue code, forked gems and interns’ earnest attempts to hack up Capistrano.  It usually gives off a strong whiff of “sloppily evolved from many 2 am patches with no code review”. This is insane.  Deploy software is the most important software you have.  Treat it that way: recruit an owner, allocate real time for development and testing, bake in metrics and track them over time.

Something that I confirmed that other conference organisers are also experiencing is last-minute ticket sales. This is something that happened with UX London this year. For most of the year, ticket sales were trickling along. Then in the last few weeks before the event we sold more tickets than we had sold in the six months previously. […] When I was in Ireland I had a chat with a friend of mine who works at the Everyman Theatre in Cork. They’re experiencing something similar. So maybe it’s not related to the tech industry specifically.

This spending of the best part of one’s life earning money in order to enjoy a questionable liberty during the least valuable part of it reminds me of the Englishman who went to India to make a fortune first, in order that he might return to England and live the life of a poet. He should have gone up the garret at once.

I would say ChatGPT (mostly the new GPT-4 model), with a lot of hand-holding and cajoling from me, wrote 60-70% of the code (PHP, Javascript, CSS, SQL) for this AMA site. And we easily did it in a third of the time it would have taken me by myself, without having to look something up on Stack Overflow every four minutes or endlessly consulting CSS and PHP reference guides or tediously writing tests, etc. etc. etc. In fact, I never would have even embarked on building this little site-let had ChatGPT not existed...I would have done something much simpler and more manual instead. And it was a blast. I had so much fun and learned so much along the way.

Nor let him ever believe that a state can always make safe choices; on the contrary, let him think that he must make only doubtful ones; because this is in the order of things, that one never tries to avoid one inconvenience without incurring another; but prudence consists of knowing how to recognize the kinds of inconveniences, and to take the least sad for good.

Consequently, it is soon recognised that they write for the sake of filling up the paper, and this is the case sometimes with the best authors...As soon as this is perceived the book should be thrown away, for time is precious. As a matter of fact, the author is cheating the reader as soon as he writes for the sake of filling up paper; because his pretext for writing is that he has something to impart...It is only the man who writes absolutely for the sake of the subject that writes anything worth writing. What an inestimable advantage it would be, if, in every branch of literature, there existed only a few but excellent books!

In 2025, Reinforcement Learning from Verifiable Rewards (RLVR) emerged as the de facto new major stage to add to this mix. By training LLMs against automatically verifiable rewards across a number of environments (e.g. think math/code puzzles), the LLMs spontaneously develop strategies that look like "reasoning" to humans - they learn to break down problem solving into intermediate calculations and they learn a number of problem solving strategies for going back and forth to figure things out (see DeepSeek R1 paper for examples).

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.

One of my insistent pleas to God and my guardian angel was that I not dream of mirrors; I recall clearly that I would keep one eye on them uneasily. I feared sometimes that they would begin to veer off from reality; other times, that I would see my face in them disfigured by strange misfortunes. I have learned that this horror is monstrously abroad in the world again...What dreadful bondage, the bondage of my face—or one of my former faces. Its odious fate makes me odious as well, but I don’t care anymore.

It’s probably a bad idea to risk paying your ransom, though — the US Treasury Dept has issued clarifying guidance that companies paying off ransomware, and all companies facilitating the payment, can be charged with sanctions violations if the bitcoins end up at North Korea or sanctioned cybercrime groups.

I've always said that setting up a web site for most folks is scary and intimidating - but myspace, with all your friends there, lends itself to a helping culture. Everyone shares how to do whatever with their circles of friends... they get by with a little help from their friends.

Submitting a paper with a "hidden" prompt is scientific misconduct if that prompt is intended to obtain a favorable review from an LLM. The inclusion of such a prompt is an attempt to subvert the peer-review process. Although ICML 2025 reviewers are forbidden from using LLMs to produce their reviews of paper submissions, this fact does not excuse the attempted subversion. (For an analogous example, consider that an author who tries to bribe a reviewer for a favorable review is engaging in misconduct even though the reviewer is not supposed to accept bribes.) Note that this use of hidden prompts is distinct from those intended to detect if LLMs are being used by reviewers; the latter is an acceptable use of hidden prompts.