AI Engineer Code Summit · Tue 10 Nov 2026 · Continental 1–3

Too much, all at once

Scott Tolinski surveyed more than 4,000 developers about how AI coding is affecting their mental health. The answers describe people who can suddenly do far more than they can hold. Read with the notes taken in the room, the talk asks one question: is this a new world people will learn to absorb, or one that burns them out first?

Scott Tolinski, Syntax · 24 min · every slide in the talk is a 3D scene built in three.js · Links and further reading
The pull

A manic loop

The first thing the survey found is that people can't stop. 52 percent kept prompting past the time they meant to quit, and those who struggled to stop also reported sleeping badly. One respondent had the AI remind them to stop for meals and still ignored it. He describes the pull as a slot machine: a variable reward, like a loot box, where the next prompt might be the one that gets it right. At the extreme, he cites a developer who ended up in hospital from stress, trying to use up tokens before a model was retired.

He isn't writing about other people. Symptoms from an earlier injury (sleep loss, muscle twitches) came back when he started prompting heavily, and he ties them to stress.

“Let me just one more. Let me just one more, right?”00:04:51
“It feels like I'm buying lottery tickets with my time and attention, but it's so fun to build.”00:06:32

Evidence: transcript 00:01:23, 00:03:25, 00:04:51, 00:06:05, 00:06:32, 00:20:46

The load

Too much at once

The pressure isn't only internal. 71 percent feel they have to produce more simply because AI makes it possible, and only 195 people said never. One respondent says writing code by hand now feels like taking a break, and real breaks feel close to illegal.

Then multiply it. Running five agents means five streams of questions; he names it decision fatigue, and quotes a React developer's escalation: two agents, four, eight, infinity. The result is people who feel more productive than ever and more tired than ever.

“Obviously, it's decision fatigue.”00:09:42
“If I can run four, I can run eight. And I can run infinity”00:09:42
Read alongsideKsenia Se's Turing Post editorial, “The Anxiety of Abundance” (Nov 9, 2026), makes the same point from outside the talk: agents take the doing off your hands, but the deciding stays with you, and it doesn't scale.

Evidence: transcript 00:07:54, 00:09:42

Meaning

A crisis of meaning

Under the overwork is a harder question: what the work even is. He says developers are in a transition in their understanding of their own work, and many are mourning the loss of writing code by hand. The survey answers split. One person lost their spark, and now it's just money. Another says AI is the spark, and someone senior enough that they rarely code is “so back now.”

“And so, a lot of people are having to mourn the loss of writing code by hand”00:12:11
The reading: this is the crisis of meaning underneath the mania. The same change empties the work for one person and refills it for another.

Evidence: transcript 00:12:11

Digesting

The brain rewires

63 percent say their coding skills are diminishing. He explains it with neuroplasticity: the brain strengthens the paths it uses, and the ones it stops using fade. At first that upset him. Then he looked at what he'd forgotten (old CSS clear-fix tricks, sites that worked in Internet Explorer 5) and decided most of it no longer matters.

“So again, we start to form new pathways.”00:14:00
The reading: this is what digesting looks like. Old skills fade and new paths form, and losing them is a cost only if they were still the point.

Evidence: transcript 00:14:00

Absorbing

Building ways to absorb it

His answer is not a tool. What makes an impact is better habits and sticking to them, which he admits is easier said than done. Run fewer agents at once: a respondent says less parallelization makes for a calmer mind, and he agrees from experience. Make room for grounded time like hobbies, pottery and hiking, experiences that were nice to have and are becoming necessary. Go outside, away from screens and agents. And keep agents out of your downtime; checking them on your phone during a hockey game is where it slips.

One respondent halved their hours, with better mental health and more output. The transcript garbles the line, but the slide reads: “I do half the work, 4x the output and a lot of time for hobbies.”

“Less paralization makes for a calmer mind.”00:19:50
“Nature heals AI psychosis. Go touch grass.”00:20:46
From the notes: one agent, watched, rather than many in parallel, is the practice that stood out in the room.

Evidence: transcript 00:17:50, 00:19:50, 00:20:46, 00:21:48

The question

Digest and transcend, or burn out?

The talk doesn't settle it. On one side, he expects change: developers are mid-transition, and as the models get better some of this pressure changes. The brain rewires. On the other side, people are exhausted and holding it in. They used the survey to say how hard it is, because nobody's talking about it. And self-control alone isn't a system.

His own deck is evidence for both. A respondent says AI let them make slides out of three.js, something they'd never have tried, and this talk is built that way: every slide is a 3D model. Abundance can be delightful and still be too much.

“It's not that easy to build these systems, right?”00:21:48
The reading: the open question is whether people digest this and come out the other side, or burn out first. The talk's honest answer is that it depends on the limits people and organizations build, and nobody has those yet.

Evidence: transcript 00:04:22, 00:06:32, 00:12:11, 00:14:00, 00:19:50, 00:21:48