Michael

Yes, progress is moving faster than any one person’s bandwidth to track it or communicate it. Even inside group 2 everyone is sprinting in their own lane code gen, security, math. Individually accelerating but unable to fully convey what they’re seeing to someone else. The perception gap just widens.

Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code. But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along. So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions. TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.

Design at Shopify is having a moment. A big moment! Incredible talent, and even stronger output. So we gave it a home: Shopify.design Come explore

@benjaminsehl the 'gave it no designs and it was terrible' is basically a rite of passage at this point

I'm building a workout tracking app/personal trainer agent. Good news is the agent part it built out is great! Bad news is it did build the app… though I gave it no designs or anything and… it was terrible. Going to have to do some proper Figma designs to show it what I'm talking about. I had too much confidence it understood my vision.

At the gym working out, while I remotely build an app I’m working on for working out. I love this timeline.

@robertbeson It would be a good choice, Dawn is part of the Horizon collection.

@miwhitham I’m trying to work out if I just build off Dawn. Will check out the Horizon collection.

1/ @fffabs shipped a whole headless Shopify storefront with Claude Code in a weekend. started from scratch, full control, great write-up. I read it and thought yeah, I'm doing similar but wanted to try to get Claude to write code inside an existing theme without breaking everything.

@robertbeson Fabric was initially closest visually to where I wanted to go. It's part of Shopify's Horizon collection which has the latest theme features. Looking back, any theme in that collection could've got me here, the real work was in the customization.

@miwhitham That's great to see. One question, why did you choose fabric?

9/ @fffabs built from scratch, ended up with something very good, and custom but owns the full stack: he owns every API change, every security patch, every browser quirk from here on out. 😅 I built on Fabric so I could point Claude at the stuff unique to my store: scoring systems, confidence tiers, comparison tables driven by metafield logic and let the theme handle the rest. Both great outcomes, using same tools, with slightly different priorities.

8/ my actual workflow: pass Claude design context or theme context or both. get Liquid/CSS back. refine by hand. repeat. not everything goes through Figma. logic-heavy components are better worked out live. the skill is knowing what context to hand over and when to just override the output yourself.

friend: just open a shopify store and sell creatine me: right right right builds a knowledge graph of every supplement brand, supplier, distribution channel, clinical study, and regulatory change dating back to 1960 😅

@jonminori you had me at beets by dre

Sharing Plots for the first time 🌱 It's got some rough edges, but I'm proud of some of the details that are there — and excited to bring that same care to the rest of the app. Would love your feedback and share your garden link in the comments. There's an easter🥚hidden in the app — stick around to the end of the video for a hint. Link in the thread ⬇️

Yeah, I've felt that hollowness sometimes finishing a task with AI help, realizing my grasp feels thinner than it should. But I think it's only natural at this stage. Most people used to hand-spin yarn from raw wool or cotton every day, it was a core domestic skill. Then mechanized spinning made it obsolete almost overnight, and within a generation or two, almost nobody knew how to do it manually anymore. The convenience was real, but so was the quiet atrophy of something once essential. We're in that early, disorienting shift again AI outsources the 'spinning,' and the struggle that used to build deep competence is optional now. The hollowness is probably just the growing pains of adapting to a new baseline. I believe the key is staying intentional about when we let go of the wheel and when we keep turning it ourselves.

Anthropic's own researchers just proved that using AI to learn new skills makes you 17% worse at them. and the part nobody's reading is more important than the headline. the paper is called "How AI Impacts Skill Formation." randomized experiment. 52 professional developers. real coding tasks with a Python library none of them had used before. half got an AI assistant. half didn't. the AI group scored 17% lower on the skills evaluation. Cohen's d of 0.738, p=0.010. that's a real effect. and here's what makes it sting: the AI group wasn't even faster. no significant speed improvement. they learned less AND didn't save time. but the viral framing of "AI bad for learning" misses what actually matters in this paper. the researchers watched screen recordings of every single participant. they identified 6 distinct patterns of how people use AI when learning something new. 3 of those patterns preserved learning. 3 destroyed it. the gap between them is enormous. participants who only asked AI conceptual questions scored 86% on the evaluation. participants who delegated everything to AI scored 24%. same tool. same task. same time limit. the difference was cognitive engagement. the highest-scoring AI users actually outperformed some of the no-AI group. they asked "why does this work" instead of "write this for me." they generated code then asked follow-up questions to understand it. they used AI as a thinking partner, not a replacement for thinking. the lowest-scoring group did what most people do under deadline pressure: pasted the prompt, copied the output, moved on. they finished fastest. they learned almost nothing. and here's the finding that should concern every engineering manager alive: the biggest score gap was on debugging questions. the skill you need most when supervising AI-generated code is the exact skill that atrophies fastest when you let AI do the work. the control group made more errors during the task. they hit bugs. they struggled with async concepts. they got frustrated. and that struggle is precisely what built their understanding. errors aren't obstacles to learning. they ARE learning. removing them with AI removes the mechanism that creates competence. participants in the AI group literally said afterward they wished they'd "paid more attention" and felt "lazy." one wrote "there are still a lot of gaps in my understanding." they could feel the hollowness of having completed something without understanding it. that's not a productivity win. that's debt. this paper isn't an argument against using AI. it's an argument against using AI unconsciously. Anthropic publishing research showing their own product can inhibit skill formation is the kind of intellectual honesty the industry needs more of. the practical takeaway is simple: if you're learning something new, use AI to ask questions, not to skip the work. the struggle is the product.

Finally

One command to clone, start, and sync everything. auto-detects your stack, streams logs with colored prefixes, and scaffolds docs so AI agents can work autonomously across the whole system. workbench init → workbench up

Built a quick CLI workbench because managing truevitamin.com across ai engine, api, frontend, mobile app, and infra repo was making me lose my mind, and time. github.com/mwhitham/workb…