Gaëtan Perrault

@futurebird@sauropods.win My 3rd grader and I actually did some book binding over the summer. Basically a coptic stitch with leather straps and an exposed spine. She had a lot of fun and she has now asked about binding different books. Once you know that it's a thing you can "just make", it really changes how you view books at all. Even if it's just "fold paper and staple the middle", it really changes the sense of what it means to create these essential things.

@RuthMalan@mastodon.social We’re going to relearn some lessons the hard way, aren’t we??? Yeah, AI is going to do this to a lot of people. For better and for worse. People regularly complain that their bosses literally didn't know what they did. Bosses are going to have to figure that out (and likely discover that they can't replace the humans). Some bosses made a living on the back of tactical ambiguity. And now they're facing a tool that doesn't do that. That needs things written clearly and explicitly. Human businesses operate with a fair amount of BS as lubricant, as signaling. We now have BS machines and they're overwhelming the system. It's going to be summer hard lessons.

@stux@mstdn.social do you have an estimate of the per user cost for that?

SQLite is kind of a magic thing. It's hard to overstate how incredibly valuable and universal it is. But in a reminder of that, I opened up two tabs yesterday where people are building new stuff using the tool. @joeycastillo@mastodon.social is building an open e-reader, and features like the built-in dictionaries and the highlighting all run off the local SQLite.@b0rk@social.jvns.ca is building new Django website with SQLite and they have great details about their trials and tribulations of mixing the two I love to see all of this open stuff happening. And thanks to both of these creators for continuing to make stuff. And check out that open e-book project!

So the news item is headlined by Trump apparently securing new sources of potash. From Belarus. At less than current Canadian prices. https://www.ctvnews.ca/canada/article/us-working-on-deal-to-buy-potash-from-belarus-for-lower-price-than-canada-trump-says/ But wait, if that potash was just sitting there, available at rates below Canadian rates, why weren't you already buying that? But likewise, if you are Belarus, why are you selling this at a discount? There are 20 countries between here and there that would happily accept a discount as well. Even a cursory examination of these questions will lead one to conclude that this potash has vibes much closer to "fell off the back of a truck", than to a negotiated trade deal.

@mayintoronto@beige.party @Amgine@mamot.fr ... And yet they want the Engineer team to maintain the monstrosity that none of their peers built... At some point, it just makes rational sense to cut over to the other things that matter to you. If you got into the field because you Loved to program software, and you just spent the last 8 months having management tell you that you're not doing that anymore, then you move to building something else you also love. To be clear, this isn't just about the robots problem. The grander scope, software has stopped really growing. It has moved very heavily to corporate rent collection. That means a lack of growth potential for workers. And it means whatever work you were doing will be cut. Whatever work you're supposed to do in the future is going to be cut. Even if it was clear the company needed to do that work. The corporate monopolies in Tech are so powerful, they can afford 100x more bugs before people start turning them off. It may be a 1000x, they won't know unless they try 🤡 //

@mayintoronto@beige.party @Amgine@mamot.fr I am with May on this one. This is the broken social contract. Doubled down with the failed state of modern capitalism. North America is filled with really bad companies and a dead end social contract. That man on the train has got to eat, so he'll fill out the TPS Reports, but when that's done, he's done. It's time to knit. Is this some weird "knowledge worker" transition? No, it's just the new reality. Is filling out TPS Reports going to post for him to have a home? Raise a family? Retire? Probably not. Is it going to get him a promotion? Probably not, everyone is laying off now.Is it going to feed him and part for a ball of yarn? Sure. So that's what you do. We can differentiate "Engineers" from "Knowledge Workers", but they're seeing similar stuff too. The career ladder has stopped growing. The robots are writing code with lots of bugs, but management actively doesn't care. Entire projects of code are being generated from half-baked CEO prompts... /1

@briankrebs@infosec.exchange @BLTpizza@mastodon.social This actually feels like one of those legal tests we need. AI bots don't have agency, they only have the security tokens of the humans and corporations that run them. Therefore the actions they take have to fall back to one of those. So any action taken by the bot is effectively an action taken by Open AI. That's currently my opinion, but I think that needs to become legal truth sooner rather than later.

@cwebber@social.coop ... with agents you can continue to make progress even when people on the team maneuvered themselves into situations where previously they would have needed to talk to each other. There's something really insidious about this take. Like it's treating humans as obstacles rather than partners. Which is kind of a big deal, because these are your co-workers, these are the other people your boss has paid to work with you. To make decisions with you. Or in the OP's case, are actually paying to work for you. This starts to sound (again) like the unitary CEO position. Where the correct number of employees is exactly one CEO and a bunch of text files operating robots.

@FuturisticRobert@infosec.exchange @mrotteveel@mstdn.social @briankrebs@infosec.exchange @BLTpizza@mastodon.social *Addendum": The reason I want to see this happen is so that we can clearly affirm this responsibility chain. I don't think it's particularly legally contentious, I just don't think it's well understood. I think that people need to see a public case of a human being prosecuted for the actions of their robots in order for this discussion to actually happen. Not just the public discourse, but also the individual questioning of responsibility. If a human knows that giving their keys to a robot leaves them responsible for the actions of that robot, that's going to change the mental calculus of those humans. I think most of us on Mastodon have already done that mental calculus. But we're the exception not the rule here.

Just a shout out to @plaguepoems@mastodon.social , who continues their important work documenting all of our plagues. I hope one day they can retire out of boredom, but that doesn't seem to be the way we're going.

@KatyElphinstone@mas.to A couple of years back, The Economist magazine did a deep dive on ADHD. They dove into difficult questions around diagnosis and measurement and the responses were quite disappointing. For example, different countries each had their own estimates for the percentage of the population that had ADHD. In part because each of them had different diagnostic criteria, so a UK human may qualify as ADHD in their country, but not in America. And part of that was political, the diagnosis wasn't just a statement of fact, it carried a measure of state support and needed to conform to state norms as well. When pressed about getting a "true count" of ADHD people, a diagnosing physician just balked. If the true percentage of ADHD people was over 10%, they simply didn't have the diagnostic capacity to handle that. ASD is in a similar spot. If "prevalence" is over 10%, we literally can't diagnose everybody.

RE: https://hachyderm.io/@norootcause/117009659827490822 This popped up first when I turned on my TV and it's a fun talk. My takeaways: Garbage collection transmutes memory saturation into CPU saturation Saturation is inevitable, resources are finite, they will run out"Prepare to be surprised": you can't predict all of the failures, so you need knobs you can tweak to respond to an incidentThe correct number of incidents is not zero, you need to be using the muscles A lot of memories from my days on call at Roblox.

@FuturisticRobert@infosec.exchange @mrotteveel@mstdn.social @briankrebs@infosec.exchange @BLTpizza@mastodon.social So this is a complex issue, but here's my quick summary. Humans have agency. They have IDs and bank accounts and stuff and bodies that can be regulated. If a human does something wrong, we can seize any one of those things. We have a long history of laws here. Corporations are a fictional entity that has been given some measure of agency. They also have an ID and bank accounts and physical stuff that we regulate. Along with some humans who are physically responsible for the fictional entity. But if a corporation does something wrong, we can seize those things. We can shut down the whole Corporation. We can even prosecute the humans. AI bots are fictional entities, but they don't have agency. They don't have an ID nor a bank account nor any physical stuff. We literally cannot prosecute them, we can't even ID them. The default in cases like this is to move up the chain. Like you would for a pet who causes harm. I would like to see that actually happen.

@pojntfx@mastodon.social Do you mean like the desktops on a workbench that you rent by the hour?

@pojntfx@mastodon.social we very much need those spaces. USB-C dock with a screen is a very cosy but practical version of this in 2026. Keyboard, mouse and wired internet all come along for the ride. It's not quite the "anonymous" internet cafes of the past, but most people have their pocket super computers these days. So the shared screen makes a lot of sense. It's interesting to me what these changes to an internet cafe setup also tell us how much has changed about computing and shared spaces.

@Rycaut@mastodon.social my first thought upon reading this was "R2D2 and C3PO become the main characters"... Artoo has seen some sh*t. But I think that would be a great campaign idea in general. It could easily be a comedy or a mystery. If you do the 5.5e thing, you could make the spirits effectively immortal, so if they die, they resurrect the next day, but they don't forget anything. Especially good for a mystery adventure, where they could progress the story without actually needing to win any battles. 🤔

@thirstybear@agilodon.social I kind of get the analogy, but it breaks down in a few of places. First off, a lot of professional communicators, who really care about the arson ring, keep calling it fire. The least we can do is stop calling it fire and start calling it an arson ring. We're not going to get a public backlash against fire, but we can get one against the arson ring. Secondly, this isn't as clear-cut as an arson ring, and I'm honestly not sure what harm mitigation we are supposed to put in place today. Facebook pirated a bunch of books to train their model. That was wrong, I agree. The American court system agreed. I don't believe that AI generated material qualifies for copyright. The American court system agrees. Screen Actors Guild negotiated all kinds of anti-ai measures. I would like to help stop the arson and mitigate the harms. But I'm honestly not sure what's next?

@chad@mstdn.ca I hope you've skipped eating for the last two days.

@tokyo_0@mas.to @juergen_hubert@mementomori.social have literally been a top poster on both of those sites. I would love to see Fediverse versions of them, I'm not actually sure how to structure it. Technologically, at this point it's not really a big deal. Big challenge with both of these sites is that their value primarily rests in human curation and feedback. In the early days, that came from a bunch of venture capital money and well-meaning people who were not burned out by too much work and not enough pay. How do we find those humans? How do we structure the governance so that people want to participate? So that the site can operate in the face of infinite spam and sock puppet accounts. I just don't know.

@0xabad1dea@infosec.exchange The craziest thing about this video is the moment I fully realized that the other dog is a ridgeback. "Look at the little puppies playing!" "Wait, the little one's a ridgeback?!?" "Keep a safe distance from the dogs!"

@NanoRaptor@bitbang.social @uberduck@hachyderm.io I love that last line, that is 🔥

@cmconseils@mastodon.social This is lovely. I wonder what it's like in full size?

"yes Dad, I have enough walkies... No Dad, you can't use this pillow, it belongs to me now" #DogsOfMastodon

@mcc@mastodon.social I am old and nerdy enough to remember .Beats... I don't know how I feel about that.

@blotosmetek@circumstances.run @jwz@mastodon.social @pl@cosocial.ca I love in Canada, all of my 4+ cards issued in the last few years, zero embossing.

@cturnbow@mastodon.world @chrisgeidner@journa.host any summary on the dissent arguments? Are the dissenters claiming some ignorance of the 14th amendment? Are they trying to invent new constitutional laws? I'm trying to figure out how this wasn't an absolute certainty?

@FuturisticRobert@infosec.exchange Sure, but I live in Canada and these are all multinational companies. Alphabet, OpenAI, Anthropic and others cannot be successful without customers in the EU, Canada and other rich world countries. It's true that you, in America may not benefit from such a decision, but others would. You'll notice when I talked about different types of entities having different types of agency, I did not discuss "free speech" or "freedom of association". Canada and the US have different legal takes on this already. Not every country is going to come to the same conclusions.

@atomicpoet@atomicpoet.org and it turns out if you get zero shots on goal, none of them go in... This felt like the NFL equivalent of a 3-0 Superbowl.

@stux@mstdn.social kind of both. I assume that you have both a baseline cost for "any person that has signed up" along with estimated costs for both "typical" posters and "prolific" posters. I assume it's not one number, I'm more interested in what any of those numbers look like from a sustainability perspective.

@wonkothesane@mstdn.social @stux@mstdn.social kind of what I was getting to over here. https://mstdn.ca/@gatesvp/117317510773682865

@superlinguo@scicomm.xyz @unicode@mastodon.social whoever wrote this article clearly had a lot of fun! Thanks for sharing.

@carnage4life@mas.to "To protect US AI buildout"... Of robots... A buildout that isn't really happening anyways. Which is particularly weird, because if the US actually had companies building actually useful humanoid robots, there would be a market for that stuff. Like this is one of those scenarios where the Chinese are not actually undercutting via State support. They've just literally managed to mass-produce a product that the US simply doesn't have. Which is especially telling, because the market for actual robots that could cost tens of thousands of dollars is the US. Skip humanoid robots, if you could make a stationary robot that folded laundry, the US is the place you would sell it first. Likely for $10k+. But the US doesn't have a laundry folding robot. That has nothing to do with Chinese competition. Trump is protecting against nothing.

@ai6yr@m.ai6yr.org "... Is becoming more mainstream... because we are about to give him a national interview... " This all feels like "CEO said" journalism. This notion that you somehow need to repeat what powerful people say, as if that's the thing we expect journalists to do. That said, I'm not even sure if it's a "decision to platform" so much as the modern SoP for journalists. Most journalists have become "documenters" rather than "investigators". If your job is to document what important people say, then it doesn't feel like platforming, it's just doing the work. I don't like it, but I'm not sure that it gets better while regular folks just keep tuning in.

@carnage4life@mas.to the language in this is wild. First off, they start with the AI-generated header image of a robot in a primary school class, but the article clearly discusses that this tool will be part of the senior school robotics and AI lab. Second, every reference to the robot's capabilities are all in future tense. The robot isn't "doing something", it "will do a bunch of things". This screams of legal speak for a product that doesn't actually exist yet, or work in any functional capacity. Like, the introduction of the product into a classroom is already an experiment. But they're making it sound like the capability of the machine to even operate is also an experiment. Thirdly, this thing is going to be a disaster. As much as I'm happy they're not putting the robot into a primary room, I'm scared at what's going to happen when you give this thing to a bunch of teens. Especially as a project. I used to work Trust & Safety at Roblox... This company is woefully unprepared.

@Ashedryden@xoxo.zone my partner and I have maintained that we are due up for a 1929 style crash sometime before 2029. In fact, we're actively planning our finances around such a thing. 1929 was the start of a pretty bad decade with a lot of hardship for a lot of people. But it was also the catalyst for the New Deal and a host of new government programs that put workers first. Sadly, I don't know how we get out of this without something similar.

@atomicpoet@atomicpoet.org @vfrmedia@social.tchncs.de I literally had this happen to me a couple of times on Reddit. I just go on the attack now. I'm a Top Writer on Quora I'm a top contributor on StackOverflow and related sitesI have been doing this for 20 years "I don't sound like AI, AI sounds like me because it was trained on my writing" It's amazing how quickly people back down when they realize they're being called out for lazy responses.

A big picture view of what the heck is actually going on with all of this #AI stuff. Delivered by long time public speaker and podcaster Richard Campbell. https://youtu.be/uWnUnMphmPM?si=C-exFIdJtf6wEe5z This is not the first time we've been sold AI.

Sharing a cover of a Canadian classic. https://youtu.be/oDu_AzEcSzk?si=jA4Zpcgt6IEy_A7C

@tomjennings@tldr.nettime.org 579 pages! It simultaneously feels like a lot and very little. I'm sure that more pages are written every year about a single popular JavaScript framework. That journal is an incredible time capsule of a time when you could actually capture things in a capsule 😀

@chrisgeidner@journa.host I really like this important line that you called out. Plaintiffs cannot argue before the Supreme Court that Executive Branch actors “unquestionably exercise[] executive power, and must therefore be controlled by the Chief Executive[,]” Slaughter, 609 U.S. at 27, and then here, argue that the Parties are sufficiently adverse to establish an actual case or controversy. It's good that a group is actually performing this review. I don't know that the actual ruling is worth anything these days, but it's important that it happened, if only for future judiciaries to track their historic value.

@cmconseils@mastodon.social This is so cool, I would love to see this in person

RE: https://cosocial.ca/@timbray/116942570335106875 This is not a feature of AI, it’s a feature of dysfunctional companies. This line is buried near the end, but I think it's the most central line of this whole piece. So much of what's happening is just highlighting the existing incompetency of so many companies. So much of the behavior we are seeing is an extension of existing problems. We're just adding a BS machine and too much money to the equation.

@MatWright@mastodon.social The other piece I read mentioned a significant language barrier between rescuers and three of the rescued. None of them had life jackets. It also mentioned that the captain was known within the chartering community. The other charters know that the captain's dead, but I suspect they are being interviewed separately. The run-in with fisheries also indicates that there's other paperwork to cross-reference. A bunch of non-english speakers, right near a border. All crammed into a boat, not wearing life jackets? I expect the authorities assume that something else is going on here, we are specifically getting only a subset of the information because everything else is under investigation.

We can talk about Canada needing to build 4 to 5 million housing units in the next decade. But that number only makes sense if we maintain the 12 to 15 million units that we currently have. And the vast majority of those units will not be up to spec in 2035. Insulation, HVAC systems, electrical upgrades, etc. Building 4 million units of new housing to a new building code is relatively easy compared to retrofitting 12 million units to that code. Those retrofits are far more labor intensive, and require more training and experience as you're dealing with 100 years of evolving building codes. And those retrofits are equally if not more important than the new housing. I think it's hard to overstate the amount of infrastructure debt we're going to be paying off over the next 30 years. //

@manlycoffee@techhub.social this sounds like a classic employer wish list for employers who don't know what they actually want. Like it's all buzzword soup, and it's moving faster than any rational human could claim expertise. The words don't have specific meanings. There's no licensing body for any of this stuff. And it's only weird to me, because it is red flags all around. If you are asking for keyword soup and I actually know what I'm doing, your keyword soup is an indicator that I'm now in consultant rather than an employee. If I don't know what I'm doing, but I can speak the words, I'm just going to snake oil sales my way through this, a thing I've seen actual people do.

@stephanie@ottawa.place I am a fellow manager of spreadsheets for our chronic illnesses. It's really good that you have done this, it's hugely valuable. But I think it's also a fundamental failing of our Canadian medical system that we end up needing to do this. You often have to go out of your way to get copies of your own data, and then you end up needing to carry it around because doctors don't actually look at data that does exist. Everyone should have access to their own version of the spreadsheet. Organized and compiled by professionals to meet everyone's needs. The fact that doctors are surprised about your spreadsheets rather than embarrassed that they're not providing it says a lot.

@mayintoronto@beige.party @PhoenixSerenity@beige.party @kaosk@mastodon.social so cool to see the Mastodonians crossing the streams