John W. Jamieson
Professor, Canada Research Chair in Marine Geology · University of Ottawa
Worst fileGlobal database of hydrothermal vent fields
Five days of learning from each other, arguing about AI, and dancing outside at −25°C. Basically a reunion with homework.
Hey, it's Sam.
We're scattered all over the place and do wildly different things. That's exactly why I want you there.
Everyone's talking about AI. Hardly anyone is talking about it with people outside their own bubble. So I'm stealing five days in February to get us in one place to teach each other what we know, have the AI conversation properly, and show you how good the free AI models you can run on your own laptop have gotten. No subscription, no internet, no one else's servers.
Then at night we go to Igloofest and dance on a frozen waterfront. Balance.
— Sam
Everyone knows something the rest of the room doesn't. A scientist, a teacher and a developer looking at the same problem is where the good stuff happens.
Not hype, not doom. What it's actually good at, where it falls flat, and what it's doing to how we work and think.
Free, downloadable AI models now run on a regular laptop with the Wi-Fi switched off. We'll set you up so you leave with your own.
One person defends a spicy claim about AI. Three people from totally different fields try to take it apart. Then the room votes with red and green cards. Get a head start: pick a side now.
AI writing helpers are making us worse at thinking for ourselves.
Most AI products are a thin wrapper that the next big model update will wipe out.
If a human expert can't make sense of a messy file, no AI should touch it.
Short, informal sessions run by us. How a geologist reads a rock core. How a teacher spots AI-written homework. How to get AI running on your own laptop. You'll learn things you'd never sign up for.
Bring the ugliest real file from your work. A chaotic spreadsheet, a scribbled field notebook, a log nobody understands. We throw it at AI live and see what actually happens.
Mixed teams get three hours to build an AI helper for scientists stuck at an Arctic camp with no internet. Then the other teams try to break it. Very on theme.
Anyone can raise a hand and say “explain that like I'm a regular person.” Buzzwords get called out. Nobody gets to hide behind jargon, especially the tech people.
Everyone's one of four types. We mix them on purpose, because the best arguments happen between people who see the world differently.
Tap the one that sounds most like you.
Three nights at Place des Festivals on the Zibi waterfront in Gatineau. Snowsuits, light shows, a crowd that's somehow warmer than you'd think. Thursday, Friday and Saturday, after a long nap.
No. Honestly, it works better if you're not. Half the point is getting people who know AI in a room with people who know everything else. If you can say “that's not how it works in my field,” you're qualified.
No slides, ever. Come with something you know well or a problem you're stuck on. If you want to teach a short session or take the Hot Seat, put your hand up. If not, heckle responsibly.
A conference where the attendees are the program. There's a loose plan, but what gets talked about comes from the people in the room. Think group project, not lecture hall.
Bring whatever you've got. A recent laptop with lots of memory is a bonus for running AI locally, but we'll share machines. I'll send simple setup instructions before the trip.
It does. Bring your valentine, or bring them chocolate after. I've chosen not to have an opinion on how you explain this one.
Expect −15°C to −25°C at night by the river. That's why every afternoon has a 3-hour break to rest and dry your gear. It's also exactly why it's memorable.
Email me and I'll send details as they firm up: where we're staying, costs, and how to get set up before the trip.
Subject line: “Frosty Boot Up: I'm in.” Bonus points if you tell me your type and the worst file you're bringing.
Every El Niño February in Ottawa since 1973, with its average high, average low and coldest night. For comparison, the long-run February average low is −13.0°C.
| Group | Low, as a bar | Avg high | Avg low |
|---|---|---|---|
| Very strong | −1.2 | −10.7 | |
| Strong | −3.1 | −11.5 | |
| Moderate | −5.3 | −15.3 | |
| Weak | −5.7 | −14.9 | |
| All El Niño | −4.2 | −13.3 | |
| Non-El Niño | −3.4 | −12.8 | |
| All years | −3.6 | −13.0 |
All figures °C. A longer bar is a colder night.
Lows follow the same pattern as the highs. Very strong events run about 2.3° milder than normal. Weak and moderate events run about 2° colder.
Three of the five mildest-night Februaries on record were El Niño years: 2024, 1998 and 2010.
A mild month doesn't mean a mild cold snap. 2016 averaged near normal but still hit −29°C, and every very-strong year dropped to −22°C or lower. Pack for −25°C either way.
| Feb | Strength | Avg high | Avg low | Coldest night |
|---|---|---|---|---|
| 1983 | Very strong | −2.1 | −11.3 | −24 |
| 1998 | Very strong | +1.2 | −7.8 | −22 |
| 2016 | Very strong | −2.6 | −12.9 | −29 |
| 1973 | Strong | −5.8 | −15.9 | −27 |
| 1988 | Strong | −4.0 | −13.0 | −24 |
| 1992 | Strong | −5.2 | −13.3 | −29 |
| 2010 | Strong | −1.9 | −7.8 | −17 |
| 2024 | Strong | +1.2 | −7.3 | −16 |
| 1987 | Moderate | −4.4 | −14.8 | −27 |
| 1995 | Moderate | −5.2 | −14.1 | −29 |
| 2003 | Moderate | −6.4 | −17.1 | −29 |
| 1977 | Weak | −4.0 | −11.5 | −22 |
| 1978 | Weak | −6.7 | −16.2 | −26 |
| 1980 | Weak | −5.5 | −13.6 | −22 |
| 2005 | Weak | −1.6 | −11.7 | −21 |
| 2007 | Weak | −6.4 | −15.8 | −24 |
| 2015 | Weak | −10.9 | −21.1 | −28 |
| 2019 | Weak | −4.8 | −14.3 | −24 |
All figures °C. Coldest nights of −27 or lower are in pink.
See who's in below. The mix of types is the point: the best arguments happen between people who see the world differently.
Professor, Canada Research Chair in Marine Geology · University of Ottawa
Worst fileGlobal database of hydrothermal vent fields
Cloud Software Architect · Edmonton
Worst filePatching logs