Innovation: Weekly Summary (October 06-12, 2025)

A week of arguing with the future

This week felt like walking through a busy market where every stall shouts a different version of the same promise: new things, better things, smarter things. Some stalls are hawking clever tools, some are selling hard lessons about old systems, and a couple are quietly offering maps to places people havent gone yet. I would describe them as a scatter of restless voices, all circling around the same idea — innovation — but from very different angles. To me, it feels like listening to a pub full of long conversations, where everyone’s a bit tipsy on optimism and worry at once.

I’ll try to pull threads here. I’d say there are five main conversations running through the posts: how we teach and think, how money and markets shape invention, the role of tools (especially AI), making and supply, and the stubborn human trait — curiosity — that keeps the whole mess moving. I’ll poke at each, point to pieces that made me pause, and toss in some plain analogies so you can picture the scene without needing a PhD in startup culture.

How we teach thinking (and why it matters)

A pair of posts this week nagged at the same nerve: education is not producing the kind of thinking people need anymore. Phil McKinney has two entries that bookend this worry — one bluntly titled I Told the Department of Education Their Graduates Were Useless and another curating Top 10 Thinking Skills & Methods. The first is a bit of a wake-up slap. The author goes back to a meeting in 2009 and says schools trained test-takers, not thinkers. He traces his own learning through tight constraints — the sort of learning you get when you cant afford to mess about — and argues that real thinking should be embedded across subjects, not taught as a separate elective. The piece gets sharp when it points a finger at AI tools like ChatGPT: they can smooth out struggle, and that smoothing is not always a good thing if struggle is where thinking gets forged.

The follow-up list post feels like a toolbox. It’s practical, with episodes and links, not just a rant. It’s the difference between hearing about a power drill and being handed one with instructions. You can almost imagine someone wiring those thinking methods into a school term like you’d fit new lights in a flat: a bit messy at first, but useful.

I would describe these pieces as insistent. They keep circling back to the same point — thinking skills are habit and craft. Teaching them the way we teach algebra, in isolation and by rote, is like teaching swimming on paper. To me, it feels like watching parents teach kids to ride a bike by giving them a lecture about balance. You need the wobble.

And then theres the dark twin: AI. Several posts touch it. One shows AI doing novel math; another has a Google VP talking about Gemini and NotebookLM as practical, blunt tools. So the worry is immediate. If schools surrender problem-making and problem-solving to polished LLMs, they might be creating generations who know where answers live but dont know how to search the house. Or how to build the house.

Money, markets, and the casino of venture

Money shows its face a lot. Some posts are angry about what money does to ambition. Credistick offers a neat, angry metaphor: the modern venture scene feels less like an engine room and more like a casino. Investors pick simple metrics — ARR, growth curves — like roulette numbers and bet, while the operators of the funds skim the chips. I’d say that piece is the oldest gripe in a new outfit. It’s basically: incentives change behavior. If you reward fast, noisy growth, you get fast, noisy companies.

That complaint hangs next to Christopher J Feola’s Business Suicide is an ugly thing. He charts newspapers that failed to adapt. Reading him, I kept picturing a shop that refused to fit electricity because the gaslamp worked ‘fine’. It’s a grim kind of stubbornness. Feola’s point is messy and human — executives who cling to yesterday’s model take their companies down with them. He points out rare survivors that reinvented their playbook, and those examples feel more like careful patchwork than magic.

Todd Gagne’s The $200 Million Mistake: Why Your Business Plan Is Already Wrong moves the conversation closer to tactics. He argues for effectual thinking — start with what you have and improvise. There’s a charm to it. Instead of writing a five-year plan on a napkin and weeping when reality steps on it, Gagne says use the 1/3 rule, the Constraint Canvas, and treat resources as relationships rather than inputs. That’s not sexy, but it’s useful. To me, it feels like cooking with whats in the fridge instead of ordering a perfect recipe kit and expecting it to arrive.

These posts read in chorus: the financial system, and the mental models it favors, steer innovation more than any solitary genius. You get the creative work you pay for. If investors want one-night stands with unicorns, thats what the market will birth.

Frontier ideas, blue-sky talk, and where new thinking shows up

There’s a nice distinction drawn by Angadh Nanjangud about what counts as frontier ideas versus blue-sky ideas. Frontier ideas are new, but they live near the shore — anchored to current knowledge and engineering constraints. Blue-sky ideas are the ones you toss on a mood board and maybe keep for late-night dreaming. Angadh says the problem is academic publishing prefers narrow technical studies, so frontier synthesis — big, cross-disciplinary design ideas — gets blurred out.

He suggests personal blogs and new magazines as better places for this kind of thinking. That bit resonated. I’d say blogs are the neighborhood garages where people tinker with prototypes, while journals are the car showrooms that prefer polished models. Frontier ideas need garages.

There’s also a recurring question: where does synthesis happen? Not often in formal papers. More likely in messy essays, talks, and personal feeds. The practical nudge from this week is: if youre after innovation that is weird but buildable, read blogs and odd magazines. Thats where people stitch ideas together across fields, using parts no one else saw fit to connect.

AI doing math and the tool conversation

Something loud happened in the AI corner this week. A piece on an unexpected mathematical proof — where GPT-5 produced a novel proof in convex optimization verified by a mathematician — made people blink. Political Calculations writes about it and frames it as a turning point: an AI not just reciting math but contributing new work.

Then theres the Google angle. Peter Yang interviews Josh Woodward, the VP of Gemini, and gives a feel for the startup-like hustle inside a giant. They talk speed, prototypes, and user feedback. NotebookLM gets a shout-out as a practical way to work with knowledge, and Woodward predicts chatbots that move beyond typing. Those two pieces sit side-by-side like two ways of holding the same match: one shows the flame producing a new pattern, the other shows the tinderbox being tuned to burn better.

I’d say the mood is curious but cautious. There’s excitement about tools that genuinely extend human reach — GPT discovering a proof is like finding an unexpected knot in the wood you’re carving — but there’s also worry. Back to the education piece: if tools do the hard work of thinking, will people get rusty? And the venture piece asks: if AI makes rapid progress, which companies benefit, and who gets paid?

The human reaction is messy. Some sound giddy. Some are nervous. Some want to weaponize the tech for profit. You can almost hear a landlord shouting about rent control in the same room where a teenager’s building a robot.

Makers, parts, and the small-economy revolution

A short, funny post by an anonymous author called Humanity is united, today, because everyone on earth needs to get some tiny little part that they’re… reads like a love note to tinkering. The author describes the absurd ways people make parts — 3D-printing, bodging, and cannibalizing things — and argues that giving people the skills to make their own small parts would change a lot.

This struck a chord with the supply chain pieces. James Wang’s Shortages Create Opportunity talks about how scarcity — like China restricting rare earths — pushes markets to innovate. He arms the theory with history: whale oil to kerosene, OPEC and the shifts that followed. He says short-term pain is real, but in the long run inventors make substitutes or new pathways. That’s the optimistic sibling to the do-it-yourself small-parts essay. When you can’t get a bolt shipped, you learn to file one.

I’d say these two pieces together feel like a call to small-scale resilience. It’s like learning to sew after your coat rips on a rainy day. You could go to the shop, or you could patch it and learn a new skill. The latter takes time but leaves you with something you can use next week when something else snaps.

Products, customers, and the art of listening

Product folks had their say too. Lewis C. Lin wrote about answering the predictable interview question What’s Your Favorite Product? using Spotify as a case study. The post is useful because it forces you to tell a story instead of making a list. He suggests showing why the product matters, what threatens it, and how an innovation could keep it alive. It’s a simple practice that forces clarity.

Then there’s Agent: Customer Development Coach for Feedback by Interjected Future — very tactical. It reads like a coach sitting next to you while you talk to customers, pointing out which questions land and which shut things down. Lots of folks forget to distinguish urgency from like. A person saying they like a feature is different from someone saying they’ll trade their lunch for it. This coach focuses on those differences. Small moves, big differences.

Pair those with Todd Gagne’s effectual thinking and you get a pattern: start with real people, real frictions, and build outward. Don’t invest in beautiful strategy documents before you talk to the folks who’ll actually use the thing. It sounds obvious, but the posts show it keeps getting missed.

Culture in large tech firms and the myth of the lone engineer-CEO

There’s a neat counter-argument in Dave Friedman’s You don't understand Tim Cook or Apple. He pushes back against the myth that Apple lost its soul because its CEO isnt an engineer. Friedman argues Cook optimized the industrial machinery — supply chain, ecosystem profitability — and that consolidation makes sustaining startup-like innovation hard for giant firms. The piece is almost consolatory: it reframes what many call decay as an industrial lifecycle.

I’d say that argument is both comforting and a bit troubling. Comforting because it explains why big firms look different today. Troubling because it makes you wonder whether our structures are preventing the wild, risky bets that create fundamental shifts. Is a company the size of Apple even the right place to expect moonshots? To me, it feels like expecting a cruise ship to do the twists a speedboat can manage. Different vessels.

Museums, funding, and public narratives about invention

A small, different note was Ian Mansfield’s report that India’s Serum Institute will fund the Science Museum gallery refurbishment. It’s a reminder that innovation isnt just product launches and VC rounds; it’s also what we choose to show in public museums. The gallery will become the Ages of Invention: The Serum Institute Gallery. A timeline of 250 years of innovation has a kind of public pedagogy: what we put in glass cases tells the next generation what mattered. This donation raises familiar questions about influence, but also does something simple — it promises an inspirational space for kids who’ll tinker in 2040.

I’d say this is the quieter form of innovation: curating memory and inspiration. It’s less flashy, but sometimes that “less flashy” is the map that guides new inventors.

Curiosity, systems thinking, and the personality of innovation

Michael Woudenberg’s Insatiable Curiosity is the emotional backbone to a lot of these posts. He frames curiosity as a systems skill that feeds everything else. He points to Jeff Bezos and Plato as reminders that curiosity needs humility and commitment. The essay reads like the best kind of mentor talk — not telling you what to build, but nudging you into the habit of asking better questions.

That’s important because curiosity is where frontier ideas and effectual thinking meet. You can teach methods and you can tweak incentives, but curiosity is the grease that keeps gears moving. Without it the machinery runs cold.

The smell of patterns: what agreed and what clashed

There are patterns that kept repeating like a chorus.

• Education and tools: lots of folks worry that tools will erode thinking unless we teach thinking differently. The anxiety around AI and schooling is specific and practical, not just doomscrolling. Phil’s posts and the AI proof piece sit close together: one warns about losing struggle, the other shows AI doing surprising, creative work. They rub against each other.

• Incentives shape innovation: Credistick, Feola, and Gagne are all saying similar things in different registers. Money and incentives steer behavior. You can dress it up with new frameworks or call it casino culture, but you end up with the same mechanic: pay for a thing and you get that thing.

• Making versus buying: the maker essay and Wang’s shortages piece both point to a long-run resilience that comes from making. In the short run its pain. In the long run its substitution and adaptation. You can almost taste the timeline: immediate irritation, then slow reinvention.

• Where synthesis happens: Angadh and others emphasize that the best new ideas are stitched from many places, and that formal venues often miss them. The place for messy, cross-field thinking is the blog, the small magazine, the kitchen table.

There were also frictions. The pro-AI proofs and Google VP optimism conflict with the education anxiety. The venture critique collides with the Google hustle piece — one voice says the system is a casino, the other says push faster, ship, learn. Those are different answers to the same problem: how to get value from new tech without destroying the conditions that made it possible.

Tiny examples, big lessons

Some small moments stuck with me like labels on jars. The customer development coach calling out the difference between like and urgency. The newsroom execs who treated digital as a nuisance instead of a new town square. The math proof that wasnt a human trick but an AI contribution. The maker who cant find a part and decides to print one. Each of those is simple. Each of those folds into a larger question: do we build systems that push folks to adapt, or do our systems reward staying still until its too late?

If you like the sound of any of those jars, jump to the original posts. Read the practical notes from Todd Gagne if you run a small team. Read Phil McKinney for a brisk shove about schooling. Read the Credistick essay if you want to be annoyed at your fund manager and feel justified about it for a bit. If you want cheering and a sense that someone is trying things inside the big machine, Josh Woodward’s interview about Gemini is good company.

Little detours that loop back

I keep thinking about a backyard barbecue. Some bits of the week are the charcoal — slow, patient, the background heat of systems and institutions. Some are the quick, bright sparks — AI wins and quirky maker hacks. Some are the conversations you have while waiting for the sausages to brown — policy, incentives, that sort of thing. You need all of it. Too much focus on the sparks and you miss the slow burners. Too much focus on the burners and you forget why anyone invited you to the party.

There’s a mild repetition across posts because the problem is persistent. People keep returning to the same tension: speed versus craft, incentives versus curiosity, tools versus habit. It’s like watching a football match where the same players keep getting the ball.

Where to poke next

If youre carrying a notebook, here are three small places to look based on the week.

• Try swapping one planning ritual for an effectual one. Make a Constraint Canvas, cook with whats in the fridge. See how quickly your assumptions change.

• If you teach or mentor, don’t add a thinking class. Try embedding thinking exercises in everyday lessons. Make the wobble part of the ride.

• If you work with AI tools, test them as collaborators, not replacements. Ask them to propose at least two bogus solutions and one plausible one. The bad answers tell you where your model fails.

These are tiny experiments. Theyre cheap. Theyre the kind of thing that feels like practicing scales on a piano — dull at first, useful later.

If any of this sounds like something you want to chew on alone, the original posts have the meat. Theyre more detailed, and some of them are practical enough to try this afternoon. If not, well, keep watching the stalls. There will be new vendors by next week, and some of them will be selling the same old promises but wrapped in a newer label. That always keeps things interesting.