You're NOT Being Left Behind in AI - Freestyle Friday (May 1, 2026)
TL;DR
Joe Reis’s core message is that AI FOMO is mostly manufactured — after speaking at DeepLearning.AI’s Dev and AI event in San Francisco with roughly 3,000 attendees, he says the real risk isn’t failing to “token max,” but getting trapped in panic about every new model and tool.
He’s intentionally moving from “token maxing” to “token minimizing” — on that SF trip he brought only a reMarkable tablet, a phone, and a hardcover book, arguing that his best work comes from slow, deep thinking rather than nonstop agent-driven iteration.
Agent speed can create a graveyard of half-remembered projects — he describes spinning up Claude Code, Codex, or IDE agents, walking away, and later forgetting the original idea, which feels productive but often turns into what he calls fake productivity.
The durable moat is still fundamentals and domain depth — Reis says people get left behind faster by weak reading, math, communication, sales, and negotiation skills than by missing the latest model release, because deep problem understanding is what agents can’t easily copy.
AI should help you build bigger things, not more incremental fluff — he contrasts superficial output with the long, deep work behind his own data modeling book and argues the world needs “what’s next,” not endless clever-but-incremental projects.
His practical advice is basically Pareto over panic — don’t chase every point solution or version bump; use the 20% of tools and workflows that get you 80% of the value, then go live at human speed instead of spending 14 hours a day glued to a screen.
The Breakdown
A conference panel turns into a bigger question
Joe opens from Salt Lake City, fresh off a panel at DeepLearning.AI’s Dev and AI event in San Francisco, where around 3,000 people showed up and a lot of the room’s energy was basically: do developers still have a future? The panel’s answer was “yes, but…,” and that “but” sends him into the real topic of the episode: the fear that if you’re not using AI at full blast every second, you’re toast.
The problem with “token maxing” as a life strategy
He riffs on the whole “maxing” culture — token maxing, looks maxing, all of it — and says the mainstream AI narrative is basically move faster, use every tool, stay on the bleeding edge. On the panel, alongside leaders from Replit, Landing AI, Oracle, and moderator Silicon Valley Girl, he pushed the opposite idea: in some cases, he’s trying to token minimize, not maximize.
The no-laptop San Francisco experiment
His example is wonderfully concrete: for the SF trip, he brought a larger reMarkable tablet, his phone, and a hardcover book — no laptop at all. He says that felt liberating because he’d been burning himself out with the feeling that if no agent was running, nothing useful was happening, when in reality his work depends on “slow but deep cycles,” not rapid-fire shallow output.
Agents are useful — and also very good at creating fake productivity
Joe isn’t anti-agent at all; he says he has workflows running in the background of both his business and life, and he thinks that’s awesome. But he also says many of us now have a graveyard of projects built by Claude Code, Codex, or whatever IDE agent we launched and then forgot about — like running faster on a treadmill or hamster wheel without actually going anywhere.
Tristan Handy, screen-light living, and doing things at human speed
He brings up a conversation with Tristan Handy, co-founder and CEO of DBT Labs, who told him we’ll probably look back at this token-maxing era next year and ask, “what the hell were we doing?” Joe says walks with just an Apple Watch, leaving the phone behind, help him think more clearly, and he likes Tristan’s home setup too: less screen time, more voice-mode ChatGPT if needed.
AI promised freedom, but everyone looks more glued to work
Joe points to the irony that AI is supposed to save us time, invoking John Maynard Keynes’s 1930 prediction that future generations might work around 15 hours a week. Instead, he says, people are working 15 hours a day, hooked on the novelty of speed and output, even as companies make usage limits tighter and the tools more expensive.
What actually matters: real problems, real moats, real skills
From there he gets sharper: stop building incremental solutions and clever little demos just because AI makes that easy. If you had infinite employees — his analogy for agents — but were still a bad manager, the outcome would still be mediocre; the real edge comes from understanding a domain deeply enough that your insight becomes the moat.
You’re probably not being left behind — unless you neglect the basics
His closing point is the most direct: the things that truly leave people behind are weak fundamentals — reading, math, communication, selling, negotiating — not missing the newest model release. He tells viewers to ignore the “artists and grifters,” trust the Pareto rule, use the 20% that gets 80% of the value, and remember his central reassurance: you’re going to be fine.