Simone Cicero

If #AI agents can instantly compare, switch, and optimize across services in real-time, then brand, network effects, and even embedded integrations lose power. The classic strategies for retaining users—whether through a superior product, first-mover advantage, or viral loops—could become temporary at best, requiring companies to constantly re-earn their relevance.

AI adoption plans count hours saved and costs cut. They skip the harder question: once organizations no longer need people to produce, what are people in them for? New Through the Boundary, on the possible, the necessary, and care: through-the-boundary.simonec…

Teams now build shared capabilities faster than the organization can say who owns them.

In knowledge industries

An organization can become increasingly capable of executing decisions without developing a corresponding capacity to deliberate on goals, priorities, and responsibilities.

The antidote to performative capture is not individual resilience or wellness, but collective stewardship of the conditions under which work happens.

"The book is wrong," I was surprised to hear @jnd1er referring to The Design of Everyday Things. That 1988 book is seminal for most of the designers I know. It brought the idea of affordances into everyday design practice. He still teaches it, he told me, but he acknowledged that it leaves out how we destroy the world while creating great products. That habit of questioning his own work runs through episode #147 of Boundaryless Conversations Podcast. And of course we started with AI. Don describes it as the first device that is "a generalist and a specialist at the same time": a designer can prototype code or a campaign and hand it off to specialists, and the boundaries between professions weaken. We pushed the idea further. If organizations are entering a post-managerial era, are we also entering a post-design one? ...where the people who live a problem design for it themselves? Don somewhat disagreed. In his view, AI gives designers more power, provided they learn the language of the business they work in. We also talked about why top-down design principles such as the UN Sustainable Development Goals changed so little, and how this new age of design-care can happen outside the market, in cooperatives and community projects that never needed a solid business model. Overall, it was an honor to host a legend like Don on our podcast. Full episode, transcript and references: blss.io/POD147

Ultimamente sento la mancanza di spazi per stare insieme senza dover produrre, performare, organizzare qualcosa. Ho scritto una riflessione sulla solitudine, sulla convivialità e sul bisogno di tornare a prenderci cura della vita comune. E forse anche un invito a trovare persone con cui sperimentare. samuraicommunity.substack.co…

Most AI programs start by asking how the organization should integrate AI. I would ask that last. First, what the technology means for your markets, supply chains, and ecosystems, and for your value thesis. Then, rewrite the vision as a thesis on human development: what does it mean to be human in this organization when AIs take over productive work? Only then, practices held by communities of practice, because they change every few weeks.

What if humans don't want to be in the loop?

BTW, if we go with pneumonic plague, we're fumbling it spectacularly - if we ever have to have a plague pandemic, let's do bubonic at least

Kinship is becoming a major business driver

Ditto nitter.cf/meedabyte/status/21080…

The figures @giansegato recently published on what Sonnet 5.5 can do, and at what cost, suggest that reasoning about architectures and systems is already becoming a cheap commodity. Defining our contribution as ever higher abstraction therefore commits us to a capability race we will keep losing, because each model release moves the frontier again. What stays scarce, and where our contribution belongs, is implication: capital, infrastructure, operational responsibility, and relationships with those who bear the consequences of a decision. through-the-boundary.simonec…

I'm concerned that what people describe as a successful implementation of an Al strategy in the organisation increasingly looks like a panopticon.

This working paper from Chicago and Stanford looked at 298,915 U.S. venture-funded startups from 1976 to 2020 and found something that inverts the usual advice. It compares three strategies: a) invent a new technology, b) swap one component inside an existing system, or c) wire together mature modules from distant parts of the economy into a configuration nobody had tried. The third one wins, and not narrowly. Moving from the lowest to the highest level of architectural recombination multiplies the odds of an IPO or a high-priced acquisition by roughly nine, cuts the odds of closure by 94%, and gets a venture from seed to early-round funding almost three years sooner. At the same time, the other two strategies mostly raise the odds of closure. In other words, for most startups, architecture beats invention. Since Henderson and Clark’s 1990 paper, this kind of recombination has been considered the riskiest move for an incumbent, which is why it gets left to entrants. This study says it is also the safest one for those entrants. That inversion, and what it means for companies that are no longer entrants, is what I unpacked in the fifth edition of Through The Boundary → through-the-boundary.simonec…

It's much easier to use LLMs to progressively complicate an artifact than to progressively simplify it.

works with documents too

You have to watch it!