Category: Web/Tech
What should I ask Moxie Marlinspike?
Yes I will be doing a Conversation with him, live at the Roots of Progress event next week. From Wikipedia:
Moxie Marlinspike is an American entrepreneur, cryptographer, and computer security researcher. Marlinspike is the creator of Signal, co-founder of the Signal Technology Foundation, and served as the first CEO of Signal Messenger LLC. He is also a co-author of the Signal Protocol encryption used by Signal, WhatsApp, Google Messages, Facebook Messenger and Skype.
There is much more at the link, for instance he is also an anarchist of some kind or another. So what should I ask him?
The polity that is Singapore
Police in Singapore have charged a man who is accused of posting an AI-generated image of a saltwater crocodile in a popular reservoir.
Ye Lin was charged with communicating a false message and obstructing the course of justice for allegedly deleting the picture and the application he used.
The fake image caused public concern, authorities allege. The national water agency suspended its work at the city-state’s largest reservoir for two days last month after receiving information that a crocodile had been spotted.
Here is the full story, via Kyle.
Don’t let AI make you dumber
That is the topic of my latest Free Press column, here is one excerpt:
I do not think the skeptics would put it this way, but as I read Conti, I find he has a pretty bleak fundamental view of humanity. Are we all really just looking to veg out and abandon curiosity and inquiry, at least once the machines have taken care of both the basic functions of life and certain higher aims such as scientific research? I think some people are like that—indeed you might say many people—but it does not reflect what I take to be the general human condition.
If I look at most people who might fit into the “middle class” when it comes to intellectual pursuits or educational status, I observe they have a lot of strong interests. This might play with their pets, improve their performance at sports, or learn how to cook better. You do not have to identify those preferences with “the new Athens” or “the next Mozart” to think they are perfectly good and noble ways for people to spend their time.
Most of us want to do something interesting and stimulating with our leisure time, and if we do not, it is often because our jobs are so busy and stressful that we just wish to decompress. Of course, in this radical vision of our AI future, fewer jobs will be so all-consuming and so more of us will use vacations and leisure time to explore and learn rather than to just sit on the beach scrolling our phones. And to the extent some jobs do remain hectic, or become even more so (such as cybersecurity), they will continue to be challenging and intellectually stimulating.
A related worry is that humans may feel they simply cannot compete with the AIs, and thus they might turn away from creative pursuits. It is true that I, more than ever, have given up all hope of proving new theorems in mathematical economics. But many of my intellectual and creative pursuits do not involve competition at all. For instance, I use AI to understand classical music better, asking the models questions before I sit down to listen to a piece. (Such as “which are the best recordings?” and “what should I listen for in the second movement?”) As the models get better and smarter, I am not going to be discouraged in this endeavor, as I was not “competing” with the models to see which of us knew more. Rather, I will gratefully end up much better informed about classical music—my increasing knowledge has already induced me to see more live concerts.
Recommended, and AI saved me time on the proofreading and fact-checking (not the writing!), so I could return to reading China Mieville…
The Macroeconomic Effect of AI through software engineering
We measure how artificial intelligence (AI) affects the economy through its impact on software engineering productivity. We use information from financial markets to develop a forward-looking measure that is available in real time. We estimate the sensitivity of each firm’s stock return to an AI stock market index, and how this sensitivity depends on the share of firm payroll in software engineering. We use a model to map this cross-sectional relationship into software engineering productivity gains. From November 2022 to December 2025, AI increased the market’s expected present value of software engineering productivity by the equivalent of a permanent 32.6% productivity increase. The corresponding effect on the level of GDP is 3.6% in the baseline and 6.5% when higher software engineering productivity also raises R&D productivity. By mid-2026, amid rapid progress in coding agents, the effect of AI on productivity and GDP had more than doubled relative to the end of 2025.
That is a new NBER working paper by
AI in science
Scientific progress is a key driver of economic growth and prosperity. There is great excitement- but also concerns- about the impacts of AI on science, but so far little data. We provide early insights on this from three data sources: a sample of 15 million Gemini interactions, an inventory of over 2,600 specialized AI models across disciplines, and a survey of over 600 scientists. We map these data to a new taxonomy of scientific tasks to study how scientists are using AI. Four main findings emerge. First, we find broad adoption and coverage: scientists use AI more than most other occupations. Specialized AI models have broad disciplinary coverage and are highly cited. Nearly half of the scientists surveyed report using some form of AI every day. Second, we document evidence that LLMs (proxied through Gemini usage) and specialized models act as complements—LLMs are used for general analysis, coding, and manuscript preparation, while specialized models provide domain-specific predictions, data generation and classification. Third, scientists report large productivity gains from using AI: a saving of nearly 7 hours per week, time which is primarily re-invested in more research. Finally, we show that AI is already changing the scientific process. As some stages of scientific research become easier, bottlenecks shift downstream. Scientists report an increased backlog of untested hypotheses and substantial demand for output verification. Our findings suggest that AI holds significant potential to increase scientific productivity. However, as with other sectors, its ultimate impact will be governed by complex task interdependencies and investment into the elimination of emerging bottlenecks.
That is from a new paper by Mihai Codreanu, et.al.
What should I ask Terence Tao?
Yes, I will be doing a Conversation with him. And he has a new book coming out Six Math Essentials. So what should I ask him?
Should you text more?
Here, in five waves of panel data (N = 1,966 US adults), we examined associations between life satisfaction and self-reported use of ten common social technologies measured every 3 months on a six-point frequency scale from ‘I did not use’ to ‘multiple times daily’. At this measurement level and timescale, Bayesian and frequentist random-intercept cross-lagged panel models showed little credible evidence that any social technology use predicts subsequent life satisfaction. In the reverse direction, increases in life satisfaction predicted only modest increases in (video) calling in select demographic groups. In analyses comparing different people, frequency of texting was associated with higher life satisfaction, whereas frequency of YouTube and TikTok use was associated with lower life satisfaction. Despite limited ability to detect within-person change due to temporal stability in responses, the absence of cross-lagged effects is informative: there is scant evidence of a meaningful relationship between social technology use and subsequent life satisfaction.
That is from a new Nature article by Kostadin Kushlev, Kibum Moon, Matt Motyl, Nathanael J. Fast & Juliana Schroeder. Via the excellent Kevin Lewis.
A doomsday scenario for American AI
That is the title of my latest Free Press column, here is the closing bit:
Sick and elderly Americans will go to Chinese companies for their AI-invented and AI-tested medical devices and drugs. America still will be a wealthy country, so China will charge the highest prices possible, yet prioritize Chinese citizens for treatment. Large numbers of Americans will die prematurely, at least compared to a world in which many of those innovations came from the U.S. My colleague Alex Tabarrok has coined the phrase invisible graveyard to refer to these lost lives, invisible because we do not observe the state of the world where they get treatment readily and cheaply. Over time, this invisible graveyard will swell into the many millions.
Finally, we will ask what went wrong.
The postmortem will be this. Many people panicked about the possibility of strong AI models killing us all. That fear was not based on peer-reviewed scientific research showing a high chance of doom, nor was doom indicated in any market prices of the time, including measures of risk. It was a story, just like this is a story, and it was spread on social media. The key point of the doom story was that, if America keeps the No. 1 spot in AI models, the models will be so strong they will do us all in, or lead to unimaginable catastrophes. We were too afraid to have America keep the lead, forgetting that if truly destructive AI is our fate, the doomer scenario can come from Chinese AI as well.
Today, we cannot say for sure that the AI doom scenario is false. But is it a story we wish to live by? Is belief in it a good way to protect and extend life, liberty, and the pursuit of happiness? Will it help us much, or for long, if it is Chinese AI that turns on us and does us in?
In my view, successful societies accept challenges and meet them. Solving problems, bit by bit, is the best way to ensure that we have the capabilities to meet big and truly existential risks, should those risks come along. Debating the chances of our doom, ex ante, on a highly speculative basis, is unlikely to provide the same kind of expertise and talent cultivation. It is instead more likely to demoralize and immobilize us.
So which America are we going to choose?
Recommended, do read the whole thing.
Peter Thiel on AI and tech stagnation and democracy
Lots of fresh material, one of the very best Peter Thiel outputs.
Earth fact of the day
An average of 72 percent of respondents said they felt curious, happy or excited about A.I., compared with 41 percent who felt worried, sad or angry.’@nytimes‘s Damien Cave uses @Gallup data to remind us the rest of the world digs AI.
It is 93 percent positive from China. That is from Nick Gillespie. Snap out of it you sad sacks, hope you can recognize negative emotional contagion when you see it! You do not all need to be regional thinkers. Here are the full rankings. Here is another visualization of the numbers:
Strange bedfellows, you might say…
*Fear of Data*
The author is Omri Ben-Shahar, and the subtitle is How Privacy Panic Led Tech Regulation Astray — and How to Fix It. I would describe this book as bracing, and full of substantive engagement. Basically the author wishes to give privacy considerations less weight in social decisions. Excerpt:
What is the concrete evidence for the benefits of facial recognition technology in investigation of human-trafficking crimes? I would love to have found global estimates of the magnitude — of the trafficking victims rescued through the most advanced facial recognition methods — but all I have is a collage of reports [reports are then described].
One chapter is entitled “The Futility of Personal Rights.” Agree or not, this book is full of actual arguments, so I approve.
From the Master of Industrial Organization
Believing that AI will be incorrigible leads us to proposing the wrong solutions, including ones which are likely to increase the danger than decrease it. I think that an AI pause is using the wrong method to fix an ill-posed problem. The point of a pause would be to research AI until we have a deep enough theoretical understanding as to predict what an AI will do. I think this goal will always elude us. I cannot conceive of what this deep theoretical understanding would look like; I don’t think anyone actually does. It’s a call for a stop, not a pause.
The world would not stand still, though. The companies far behind the frontier now would doubtless learn more about what it is that the frontier AI companies are doing. If cooperation were to break down, we would have restarted the race with many more competitors. This is no improvement.
Optimal liability for offensive and defensive AI
How much liability should AI providers bear when their services enable both attack and defence? Liability can improve welfare while increasing harm. Providers sell a common input to productive users, attackers and defenders. Within a defended contest, a higher common price reduces effort without changing attack success or attacker profits, saving resources and improving the target’s security payoff. Compensation weakens defence and raises attacker profits. Optimal liability balances these effects against productive exclusion. Greater competition can lower optimal liability; every such decline must end at an outcome retaining defence. With cybersecurity access fixed, monopoly can warrant partial liability but never full liability when provision is worthwhile. When guardrails preserving productive uses are available, strong competition favours universal guarding socially but encourages unilateral removal at insufficient liability. At a fixed provider count, sufficiently many productive users ensure a pure equilibrium with universal guarding under high liability. A universal-guarding requirement makes liability redundant. Under monopoly, adoption follows a unique liability threshold, while zero liability remains uniquely optimal for a range of parameters with sufficiently many productive users.
That is from a new paper by Joshua Gans.
Dartmouth Provost okie-dokie
On Aug. 26, a Semafor investigation using Pangram to determine whether recent guest columns in major newspapers were written using artificial intelligence identified a column by Schnell as “100% AI-written.” Schnell’s column in The Washington Post — titled “Universities are fighting AI cheating. But there’s a deeper problem.” — argued that “a degree should distinguish what students can do independently from what they can accomplish with AI.
”The Dartmouth subsequently reviewed publications written by Schnell before and after the November 2022 public release of ChatGPT. Pangram 4.0, the latest publicly available version of the AI detector, labeled all of Schnell’s written works prior to 2022 as “100% human-written,” while a test of nine of his works published this year returned a median “AI-written” percentage of 96%. Between 2019 and 2026, Schnell did not publish any academic works or op-eds for which he was the sole author credited, according to his online curriculum vitae.Pangram 4.0 assesses whether a work is AI-written by separating the submitted text into smaller segments, which are then identified as “human,” “AI-assisted” or “AI-generated,” according to a Pangram technical report.
…Schnell’s Sept. 18 email statement to The Dartmouth was itself “100% AI-written,” according to a Pangram test.
Here is the full story, you people crack me up…
Banning self-recursive improvement in AI models?
Some people are suggesting this, including Ezra Klein.
I do not understand how it is supposed to work. Put aside the issue of foregone innovations, let us say I seek to access an AI from overseas. Am I allowed to visit their website? So many sites have AI behind them, including Chinese open source AI. Do all of those sites get banned? And how? Firewall imposed on Americans? And we cannot put the good open source models on our hard drives? How enforced? Does the government have to read the creation logs of the model (how can they?), and then decide whether I can use that model or not?
What if Anthropic licenses its IP to an independent subsidiary in the Cayman Islands and they keep on using code to improve the AI? Or maybe only foreign nations are allowed to have the best models and the really fast rates of improvement? How are multinationals supposed to operate across borders? (Presumably you cannot just use the RSI of your foreign affiliate, but then the whole MNC is crippled and eventually rendered uncompetitive?) How are we to keep “AI sovereignty” for Americans?
If the top U.S. labs have to give up on what is supposed to be really important, they will look very “catchable.” The policy creates a big incentive for some other parts of the world to invest a lot more in compute. How safe is that? And in the short run, do we also have to stop American firms from selling their compute abroad?
Does this whole thing mean that regular companies can use AI to write code, but the main labs cannot? Solve for the equilibrium there. Or the whole code writing function goes away for everybody?
What is the political mechanism through which RSI progress gets turned back on again, and when? Is it as rational as our current debate over data centers? If the top companies are no longer building RSI models, how are they supposed to know how safe or unsafe they might be?
At current margins, are humans even capable of writing the next steps of code that are required for further progress? At whatever speed?
I do not find this to be a workable idea. I think it is “people wanting to do something that sounds reasonable and less scary,” without thinking through the practical details. I have made a related point before, but one does best with AI policy when you start with the stuff that otherwise you might put into your last two or three paragraphs of an essay (“It remains to be seen how…but we are all human beings and surely we should not give up hope…”). Even if you believe in a deal with China, and I do not, we (and others) now know competitors can catch up more quickly than we had thought, at least if they think they have a chance of winning. They might even buy some space compute from Elon.
C’mon people, I know for many of you this is a difficult take to swallow, because it feels like “we don’t care enough” if we do not take some dramatic steps. Nonetheless the reality is that, a long time ago, we made a whole series of decisions that imply…we have no other choice than to just see this one through and get it more or less right. The longer we deny that truth, the harder it will be for all of us.