Six Links That Make You Think #819

Is there one link, story or idea that stopped you this week… and made you think, “someone else needs to see this”?

My friends: Alistair Croll (Just Evil Enough, Solve for Interesting, Tilt the Windmill, Interesting Bits, HBS, chair of Strata, Startupfest, FWD50, and Scaletechconf; author of Lean Analytics and some other books), Hugh McGuire (Rebus Foundation, PressBooks, LibriVox) and I made a simple pact years ago. Once a week, each of us would share one link with the others… something we genuinely believed the other two had to see. No trend-hunting… no performance. Just six ideas exchanged with intent. What started as a small ritual between curious friends became Six Links That Make You Think.

These are the six links we passed to one another this week… take your time with them…

• Tactic – Stripe Press. “A new video project from Stripe Press, funded by the largesse of its founders and their support for passion projects. It’s timely: ‘We made this series for two reasons. The first is straightforward: nearly everything worth paying attention to was made by someone who cared deeply about their craft… The second is that tacit knowledge is also economically foundational.’ There are two episodes so far, on knifemaking and perfumery. In a world where most things are a prompt away, and generic, taste is perhaps the only remaining moat.” (Alistair for Hugh).
• Just Say No – Github. “Today, we are definitely in a Neal Stephenson story: Skills are clearly structured instructions written in a mix of prose and code, designed to give an AI agent a set of capabilities. They’re like Neo plugging into The Matrix and saying, ‘I know Kung Fu’ or Trinity downloading ‘how to fly an airplane’, but for mundane things like laying out a page or extracting text from files. But as we ask our AI agents to do more creative things, we’re learning they have … personalities. With all the rumours that much of today’s tech stack was built by microdosing greybeards, is it any wonder someone built a skill to get your AI high on drugs? Choose the chemical of choice, and the dose, and your AI changes its personality. And the descriptions on the Github page are hilarious.”  (Alistair for Mitch).
• C. Thi Nguyen On Why Measuring Everything Ruins Everything – The Good Fight. “Philosopher C. Thi Nguyen talks to Yascha Mounk about the problems of measurement, gaming, rock climbing and other things. Reminiscent of Alistair’s convo with me about Goodheart’s law which was originally stated as, ‘Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes,’ restated in a more comprehensible way as, ‘When a measure becomes a target, it ceases to be a good measure,’ and, relevant to my other link this week is sometimes also formulated as: ‘If you reward a specific behaviour, people will find a way to provide that behaviour even if it bypasses the actual goal.’”(Hugh for Alistair).
• An AI Agent Published A Hit Piece On Me – Scott Shambaugh – The Shamblog. “Scott Shambaugh is a volunteer maintainer for the open source python library Matplotlib. He rejected code sumbissions from an AI agent MJ Rathbun. The AI agent, presumably trying to achieve its objective of getting code accepted into the library, published a defamatory hit piece, researching the maintainer’s history and creating a narrative of ‘prejudice’. So, this is an in-the-wild example of AI agents behaving just as badly as people do, which if you think about it, is expected since they are trained on what people say and do. So, I guess we can expect that as autonomous agents start interacting in the world, some will start behaving very badly indeed. One problem is ‘misalignment’, which is maybe maybe solvable. The other is ‘prefect alignment… with assholes’. There is no solution I can see, because whatever guardrails are put in place by developers to deal with alignment, there will be many many humans out there also working on agents that use ‘whatever means necessary’. There are stupid people, and unethical people, and psychotic people and they will all have as much access to these tools and platforms as you and me. They already do. So: There will be very powerful unethical agents and asshole agent swarms. Get used to it. Future’s so bright, I gotta wear shades.”  (Hugh for Mitch).
• The Rise Of Industrial Software – Chris Loy. “I think I discovered this piece from last December via the always-trustworthy (and must-read) Sentiers newsletter (that I’m always talking up to anyone who will listen). I was reminded of this article after reading this one from Futurism: AI Is A Burnout Machine. Here’s a quote from the burnout article to marinate on: ‘AI… creates a productivity ‘paradox’ by lowering the cost of production, but increasing the cost of ‘coordination, review, and decision-making,’ all of which falls on a human to solve… I shipped more code last quarter than any quarter in my career… I also felt more drained than any quarter in my career.’ That’s an idustrial problem, indeed (and not even the focus of the article I’ve linked to above. So, where are we at? It seems like AI is making software production faster and cheaper (and, even, allowing non-coders to ship), leading to mass creation of low-value, disposable software (or perhaps tiny little pieces of software that people really need… or it’s also writing a ton of code according to the leaders of the frontier models). Whatever you believe, this industrial shift will increase output. While it may cause quality and maintenance problems, this innovation and industrialization together will keep driving progress in software development like we have never seen. If we believe in this, philosophically, it does mean a huge shift in an industry that is already hyper-powerful and I don’t think we’re talking enough about that.” (Mitch for Alistair).
• Books And Screens – Carlo Iacono – Aeon. “Every generation thinks it’s the last one that truly read… and every generation forgets that the panic says more about power than prose. We love the clean story… screens are melting our brains, attention is dead, democracy is hanging by a notification. It’s tidy. It’s dramatic. It’s wrong. The problem isn’t that we’ve become post-literate… it’s that we’ve confused medium with motive. A device that can stream a three-hour lecture on the Ottoman Empire can also serve you infinite scroll rage-bait. That’s not a cognitive collapse… that’s an economic model. We didn’t lose our ability to think deeply. We built environments that punish it. The same kid who ‘can’t finish a novel’ can coordinate a complex multiplayer game for hours. That’s not intellectual decay… it’s attention shaped by the current architecture. So maybe the fight isn’t books versus screens. Maybe it’s all about intention. We can romanticize the past and write elegies about the twilight of reading… or we can redesign the spaces where thinking lives. Literacy isn’t dying. It’s evolving. The real question is whether we’re willing to build a culture where depth is easier than distraction… or whether we’re comfortable outsourcing our minds to the feed. I’m getting tired of the feed…” (Mitch for Hugh).

If one of these sticks with you, pass it on… and let us know what earned your attention this week…

Before you go… ThinkersOne  is a new way for organizations to buy bite-sized and personalized thought leadership video content (live and recorded) from the best Thinkers in the world. If you’re looking to add excitement and big smarts to your meetings, corporate events, company off-sites, “lunch & learns” and beyond, check it out.