All roleplaying games have at least a dual nature. It helps to know that the company that invented the first one, Dungeons & Dragons was TSR, which stands for Tactical Studies Rules. It was a company making wargames, and D&D at its core is a squad-based fantasy wargame. Tactical combat is an important part of nearly all roleplaying games, and even more so in the computer games. The other layer of RPGs is the story layer. What is that group of fantasy heroes fighting for or against? There frequently is an archvillain or some other greater evil, and in high fantasy settings the heroes might well end up saving the world, or at least a part of it. Now pen & paper RPGs usually just stick to those two parts, but in CRPGs there is quite often also a bit of a management part to it. The group of heroes has resources, for example spells, which deplete over time and recover by resting, and you need to manage these resources. You also need to manage equipment and inventory.

Pillars of Eternity has a big world-saving kind of story, about a very peculiar type of world. In this world, reincarnation is real, and magic / technology can be used to meddle with the souls that are transferred during reincarnation. The story deals with various aspects of that, from the main characters ability to remember his previous incarnations and speak with errant souls, to the major story about what gods are in such a world of reincarnating souls. While the writing is often quite good, it has to be said that the story is very complex, very long, and somewhat esoteric. It doesn't have the "tadpole in your eye" immediate directness and urgency of Baldur's Gate 3. Much of the story of Pillars of Eternity is only presented by text without voice acting, and that involves a lot of reading. In the end you don't really know whether you are chasing the archvillain because of what he did to the world, or just because you want to get rid of your status as a "watcher", as that is driving you crazy.

The walls of text problem of Pillars of Eternity is made worse by the game having been financed by crowdfunding. This allowed backers who paid more money than others to add their own text to the game, in the form of NPCs with their own backstories, or in the form of epitaphs on graves. You quickly learn that clicking on an NPC with a gold nameplate and reading his backstory is just a waste of time, as that story was written by a backer, and doesn't really have anything to do with the story of the game. Kickstarter stretch goals also added other unnecessary ballast, like an "endless dungeon" to crawl through for no reason, or a half-baked system in which you manage a stronghold, much worse than in Pathfinder: Kingmaker.

I am playing Pillars of Eternity now, because the game this year added a turn-based mode. For me, the tactical combat is more fun in turn-based mode, as I find real-time with pause frequently confusing. Having said that, I reached the point where the tactical combat isn't that much fun anymore. To me it seems there are some rather fundamental balance problems in the system. Due to enemies damage reduction being deducted from any damage you do, weapons that deal lots of damage slowly to me appear much more powerful than weapons dealing smaller packets of damage more frequently. And because the world of Pillars of Eternity has firearms, and firearms are slow and deal lots of damage, all of my group is now equipped with firearms or heavy crossbows. While against a large group of enemies some spells like fireball are very useful, against a single enemy my mage is dealing more damage with a pistol than with a spell. And the spell is a resource that depletes, while the pistol has unlimited ammo and shots.

Another reason that the game has become a bit too easy for me is the structure of many computer roleplaying games that have a main quest and optional sidequests. The dungeon I am in is apparently meant for level 7, but my group is level 9, because I did every single available sidequest. And that is without the DLC, of which is part is designed to be played at level 5+, and which results in your group being even more overleveled for the latter part of the main story. But the sidequests to me actually were more fun that the main story, as these smaller stories are easier to understand, and often deal with interesting moral dilemmas. They are more relatable than the main story about souls and reincarnation, because the moral dilemmas are of a kind that could also occur in our world, not just in a world with a specific way souls work. Being somewhat completionist with the sidequests and exploration also led to the management part of the game having become trivial, as Pillars of Eternity allows you to take all loot everywhere into a stash with unlimited capacity, and sell it from there later. I have far more gold, gear, scrolls, potions, and materials for crafting than I will ever need.

I think I am going to somewhat rush the rest of the game, and if that doesn't work, I'll stop playing before reaching the end. I can't remember a single computer roleplaying game in which I found the grand finale big boss fight satisfying. Some games do better than others by making the end feel better from the story aspect, but the final tactical combat is almost always somewhat tedious. Which isn't surprising, as even in pen & paper Dungeons & Dragons I find character progression at lower levels a lot more interesting than the high-level stuff.

Why do big game companies make triple A games for hundreds of millions of dollars? It seems to me that the model for this was the movie industry. When I was young, cinemas still showed a mix of large blockbuster movies and smaller production, even B movies. But the economies of scale of movies made it so that the huge blockbuster movies were more profitable than the cheap movies. These days, indie movies are something only a few specialized cinemas in large cities might still show. The general movie market only has films that cost hundred of millions of dollars, and these crushed the smaller competition.

Games have very different economies of scale, especially when distributed digitally. The moment a game arrives on the Steam platform, there is nothing to prevent it from being sold millions of times at no additional cost to the developer. But there is also nothing to prevent it from being a total flop and selling just a few thousand copies. It then becomes a consideration of risk vs. reward: The two Japanese developers only risked two months of work, and if their game hadn't sold much, it would not have been a big loss for them, and maybe a fun experience. They got lucky and made millions, and after the share that Steam takes, they only need to split the money between the two of them. In comparison, 6,000 people are working on GTA 6. Now GTA 6 is extremely unlikely to flop, and will probably also sell millions. But its budget is over $1 billion, and it can't possibly reach the same return on investment as Meccha Chameleon. It can get its production cost back, and make a nice profit for the game company and its shareholders. It can't sell ten times more or hundred times more than expected, because the expectation is already selling 40 million copies, and that is approaching market saturation.

Of course any videogame of any size and budget can be a failure. The trouble with triple A games is that they aren't guaranteed to make their development money back, even if they just sell an average number of copies. That is especially bad for live service games, which are expected to not only sell well, but produce much additional revenue over the further lifetime of the game. The larger the company and the bigger the budget, the more people lose their jobs when the game just gets a lukewarm reception. Smaller studios and indie developers can more easily recover after a failure. The difference between the games industry and the movie industry is that there doesn't seem to be much correlation between the budget of a game and the probability of it being a commercial success. You can spend a hundred million dollars on making a game and sell less well than Meccha Chameleon. That is a pretty bad business proposition.

People like Elon Musk and Sam Altman for years have been promising that AI would reach "PhD level" intelligence. Now I happen to have a PhD degree. Over the course of my life, I have been producing piles of "PhD level" research work, starting from my actual PhD thesis and covering 30 years of work as a researcher in industry. But while my biased opinions about World of Warcraft have been scraped by AI and are now part of the models, my "PhD level" work isn't in those models. My PhD thesis is not available in electronic form anywhere, it only exists in paper form. My industry research is either long lost, or locked away in the computer systems of my employer, and not accessible to the public or AI scrapers. I have a number of scientific publications that are available, and my 32 patents are also public, but these are small windows into my scientific work. The bulk of my scientific work is invisible, proprietary to my employer, and not being shared. While in theory a patent protects an invention in exchange for sharing that invention, in practice there is an internal process that makes sure that the protection is maximized, and the sharing is minimized. Having done applied industrial research, I did do research studies that made my company millions in profit, but none of that will ever make it into an AI model.

I assume that this is the same for quite a lot of people. Many of us produce some written output both in our private lives, e.g. on social media, and our professional lives. But our professional output is legally owned by our employer, and guarded as proprietary and secret information. The part of as any AI model can possibly know is just what we produce publicly, in a private capacity. You can find a video on YouTube on how to change a spark plug, but not the totality of the professional knowledge of a car mechanic. And the useful information on social media is heavily diluted; already for a human it is hard to separate the useful stuff from the chaff, but an AI model just takes everything, and has notorious difficulties in separating facts from beliefs or jokes. It takes just one joking Reddit post telling you where you can stick that spark plug before the AI might end up repeating that as medical advice.

Long before ChatGPT was released, we already knew that a lot of the stuff you read on the internet is garbage. In the early days of LLM models, humans moderated the input of those models, feeding it for example digitally available books rather than unmoderated forum discussions. But with growth came the need for more and more data, and the AI companies became less and less fussy about the quality of the data being fed to the models, because they needed so much of it. If you grab terabytes of data from the internet, very little of it will be "PhD level" intelligent, and very much will be garbage. Garbage in, garbage out, is one of the oldest truths in computing.

A LLM model could probably replace me as a blogger. I have been feeding the models enough data to be able for them to simulate that part of my activity. But as they have extremely little data on my professional work, I don't see how a LLM model could do my job as a researcher. Even if there was actually some "intelligence" in artificial intelligence and those "reasoning" functions could actually reason, the models simply don't have the professional data that would allow them to do professional jobs. And no company would ever open up their proprietary company data to a public AI model. It is not about the quantity of the data, but about the quality. If I get stuck in a game, I totally trust AI to tell me how to proceed, because that sort of data is readily available. I wouldn't trust AI to engineer anything or to research anything.

I have to say that I dislike swapping companions in and out of my party repeatedly. Of course, with some foreknowledge it is possible to always choose the optimal party configuration for every hard battle. But I am more likely to swap companions because I'm going somewhere which relates to a personal quest of that companion. You will want to have Lae'zel with you when dealing with the Githyanki, because otherwise you are missing out on the specific content of that interaction, and so forth.

But from a pen & paper roleplaying point of view that companion-swapping doesn't make much sense. Characters in a tabletop RPG would only change if either a new player joins the group, or a player for some reason switches to a new character. I over 40 years of D&D I never had a player switching back to a previously played and discarded character. A new player with a new character would also simply just add to the group, not necessarily replace a previous player with a previous character. But games like BG3 or Pillars of Eternity have hard caps on the number of characters in a group, so you can't just take an extra with you for the occasion.

I tend to get used to the characters I am playing with, and interested in their stories. Thus I am reluctant to bench them, just to get to know a new character and get involved in his story. It feels like ditching a friend. So maybe a better way would be if a game just had as many companions as there are free slots in the group, but those companions could adapt their class to result in a balanced group around the main character class choice, while keeping their personal stories constant.

In general, I am well equipped against temperature extremes, as my house is built do "quasi zero energy" standards, with 36 cm thick walls in three layers and triple glazing on the windows. But after several days with daily temperatures in the mid-30s Celsius (up to 100°F), and not much cooling during the night, the rooms in my house are around 27°C (80°F) and it's getting uncomfortable.

The obvious solution, as all my American readers will immediately say, is air conditioning. A few years ago, when heat waves went from being very rare to occurring occasionally, I bought a good quality mobile air conditioning unit. It is a monoblock unit on wheels with 12,000 BTU of cooling power. At the start of the current heat wave it still managed to cool down our bedroom to 20°C before we went to bed. Enough to turn it off, fall asleep, and not get too hot by the end of the night. Right now that air conditioner is struggling. The problem is that you need to open the window to put the vent hose outside; and as much as I'm trying to seal the window around that hose, at 36°C outside I can't get the inside much cooler than 24°C. Which is an improvement to the 27°C before, but not a huge one. And if I cool my bedroom to 24°C, it gets too warm to sleep during the course of the night.

So I am considering adopting American habits and installing a fixed split air conditioner. Which might be expensive and difficult due to those thick insulated walls. In any case, while my monoblock air conditioner cost several hundred bucks, a fixed unit will be several thousand bucks with installation. On the positive side, a fixed unit with the vent going through the wall means that I wouldn't have to open the window to run the air conditioning, and I could achieve a higher difference between inside and outside temperature. In a situation where 40°C becomes possible in Belgium (the current heat record is 39.9°C) and heat waves are becoming more frequent, more intense, and longer, this investment might become necessary.

I played Pillars of Eternity in 2015. But only for 16 hours, never even finishing Act I. In 2015 I was still playing World of Warcraft, and we all know how time-consuming that can be. But as I wrote in my previous post, I have been watching a streamer playing Baldur's Gate 1, and that revived my interest in games that use this sort of graphics engine. Now I have two problems with games like Baldur's Gate 1: Real-time combat with pause isn't really my favorite, and the pre-rendered rectangular maps are often too large; that is to say, the game doesn't really have enough content for the size of the map, and you end up spending hours walking all over the map just to uncover the fog of war. Now Pillars of Eternity, having been released 17 years after Baldur's Gate, is already a lot more modern, didn't fall for the same trap, and made maps somewhat smaller than BG1. And then Obsidian Entertainment in a somewhat surprising move added turn-based mode to Pillars of Eternity this year in April.

So I decided to give Pillars of Eternity another try. To my surprise my save game from 2015 was still there due to the power of Steam cloud saves. But I prefered to start afresh. In my 2015 game I had played a rogue as my main character, for the simple reason that this was the character class most notably missing from the early companions you can pick up. In my 2026 run I went with a priest as main character because I like playing priests, and they are rather strong in Pillars of Eternity. And then I used previous knowledge as soon as I reached the first village to hire a rogue hireling to fill that lockpicking / trap disarming gap; and I skipped ahead a few zones to collect the first 4 companions: Edér the warrior, Aloth the mage, Durance the priest, and Kana the chanter. This allowed me to do the quests and the first dungeon of the starting village with a full group, which worked very well.

I'm only 8 hours into this second run. Apparently doing the main story plus side quests without the DLCs takes up to 70 hours. I don't know if I'll make it to the end game screen this time, but I'll give it a try. And then there is a still unplayed Pillars of Eternity 2 in my Steam library, so I'm unlikely to run out of 2D pre-rendered isometric map games anytime soon.

As the game predates Steam, I have absolutely no clue how many hours I played of BG1. But I did play it through, and I did play BG2 and BG3. Baldur‘s Gate 3 I played for 340 hours, and in my opinion, it is the best game of the series. I‘ll play more BG3 before I‘d come back to BG1 or 2. But the Baldur‘s Gate series is special insofar as there is this huge gap of nearly a quarter of a century between 2 and 3. Even with the enhanced editions, the technological advantage of BG3 over the previous games is huge, and the design philosophy of computer roleplaying games has also evolved in that timeframe.

Technological advancement of PC games has slowed down. A game of 2026 isn‘t necessarily much more advanced than a game from a few years ago. And game series other than Baldur‘s Gate with more sequels than 3 and shorter time between sequels have a bit of a problem: The latest sequel isn‘t necessarily considered by everybody as the best in the series. As a consequence for example right now there are more people playing Civilization 6 than Civilization 7, and more people playing Europa Universalis IV than EU5.

If you look at the currently most played games on Steam, a lot of the games in the top 50 are older games. Newer games of the same series or at least the same genre exist, but players prefer the older games. The top 3 games are Counter-Strike 2, PUBG:Battlegrounds, and Dota 2. And people busy playing older games simply often don‘t have enough time to also play all the latest games, and so they don‘t buy them.

On the one side one can argue that making every game better than all previous games is an impossible task. On the other side we all know lots of examples where a new game feels like a beta test, being full of bugs and/or balance issues. I stopped playing EU5 because every patch is still messing very much with the game systems and that often has unforeseen consequences. I can see how other people would prefer a stable and established EU4 to this mess.

The „live service“ genre of games is most affected by this, and it is rare that a new live service game can gather critical mass and survive, while people are still happily playing much older games. But while things are less visible for single-player games, Steam player counts do at least suggest that sales of newer games must be down, seeing how many people are busy playing older games of the same genre. And game companies increasingly release remakes and remasters of older games, to make at least some money from all these people preferring the older games.

I don‘t think that the games of next year can beat last year‘s game just on prettier graphics. At some point graphics are „good enough“ and stop being a major criterion when buying a new game. But while good new games are still being made and have success, these all work with original content rather than using better technology. In a world where you can‘t watch the news without hearing the word „AI“, the progress being made on video game AI providing more realistic opponents or NPCs is practically non-existent. Many of the top-selling games of 2026 could have been made a decade ago, they don‘t use any technological progress at all. And unless that changes, older games will continue to occupy a large space in this attention economy, making life difficult for new games.

Yesterday was the IPO of SpaceX. It's IPO price was $135, and over the first day of trading the share price stabilized at around $160. If you multiply the total number of existing shares in SpaceX with that share price, the company SpaceX is worth a bit over $2 trillion. That makes is roughly 10 times as valuable as the company I own shares in. But if you look at SpaceX revenues, profits, and dividends, there is no logical relation to the share price. SpaceX has no profits or dividends. It makes roughly 10 times *less* revenue than the company I own shares in. In other words, the company I own shares in has a share price to revenue ratio of slightly over 1, while SpaceX has a share price to revenue ratio of slightly over 100. SpaceX is what is politely called a growth stock. The numbers only make sense if you assume that the revenue and profits of the company will grow at a rather fantastic rate.

Are the people buying SpaceX shares all fools? No, many are very rational people. For example there are people who undersigned the IPO, got shares for $135 from SpaceX, and then directly sold them on the first day for $160. $15 pure profit per share in just a day, nothing foolish about that. IPOs these days are carefully orchestrated. SpaceX only sold 5% of its shares, to keep the supply of these shares lower than the demand, more or less guaranteeing the share price would "pop" up on the first day of trading, creating a positive narrative.

But what about the other 95% of SpaceX stock, not currently traded? Elon Musk owns around 45% of SpaceX (in a special class of shares with higher voting power, giving him around 80% of votes and thus total control). Elon probably doesn't want to sell those shares, and is actually contractually prevented from selling them for a year. That still leaves 50% of shares in the hands of other SpaceX founders, early venture capital investors, and SpaceX employees. They too are still under a contractual lockout, preventing them from selling their shares, but not for long. SpaceX has a staggered lockout, in several tranches from day 70 to day 180 after the IPO. In other words, other than Elon Musk, over the next half a year, all SpaceX shareholders become free to sell their shares.

Thus the supply of SpaceX shares is going to increase a lot between now and the end of this year. That makes the trajectory of the SpaceX share price somewhat predictable. So predictable that is has a name, the "IPO pop and drop". It is very basic economics, supply and demand. Supply of SpaceX shares is increasing over the next 6 months, demand probably by not that much, thus a drop in share price is rather likely. In fact, some people are planning to make a lot of money short-selling SpaceX shares.

Now this certainly isn't financial advice. But if you were a strong believer in growth stocks, a believer that everything Elon Musk touches will turn to gold, and you do think that one day SpaceX will reach revenue and profits that are more reasonably aligned with its market capitalization, you might still want to wait for several months before buying SpaceX stock. There are a lot of stocks that are worth more today than they were on their first day of trading, but there are very few for who the first day of trading would have been the best day to buy them. And that is much more true for growth stocks than for value stocks.

Many of the people buying SpaceX stock now don't plan to wait until the company makes a profit and they get their money back via dividends. Instead they believe in the greater fool theory of investing: It doesn't matter whether the share price of a company is justified by that company's revenue and profit; you can make a profit from an overvalued share as long as there is somebody out there who is a "greater fool" than you, and who values that company even higher than you paid for it. That is purely speculative. But it is the explanation for the valuations of cryptocurrencies, NFTs, and many growth stocks, especially in AI. You are not betting on a company reaching certain goals, you are betting on the mood of the market. That isn't investing, it is gambling. There are certainly people that made money that way. And there are people who still own a bored ape NFT, who bought Bitcoin for $100,000, or bought Pets.com shares at the IPO. In a market dominated by professional traders, how likely do you think a regular retail investor is to outsmart the bankers? If you want to get rich quick, you are probably better off betting on the outcome of soccer world cup. If you want to save money for your retirement, you should stick to boring value stocks.

While I managed to play the same civilization over the ages, that actually was never really my problem with Civ 7, and I had enjoyed other 4X games in which you didn't stay the same for the whole game. My problem with Civ 7 rather was that the second age, exploration, feels very scripted and sameish to me, regardless with which civilization or civilizations I play it. It is just the tech tree that allows you the exploration, and the whole reward structure is designed to push you into that exploration.

So between the uncanny unfamiliarity with the game and feeling that I was still following the same scripted gameplay, I stopped half way through the age of exploration and gave up. For me this wasn't the patch that fixed all of Civ 7's problems and made me come back. And when I see the concurrent player numbers, it seems that the rest of the world feels the same. After a short peak Civ 7 is back to have far fewer players than Civ 6.

Not all board games are created equal. Something as simple as a deck of cards can have various different levels of card stock and finishing, so that those cards could be flimsy and wear out quickly, or be a lot more solid and wear resistant. Tokens can be cardboard, wood, or plastic. Meeples can be cardboard standees, wooden, or plastic miniatures, with or without sundrop or paint. Chip Theory Games is famous for making games that are completely waterproof: Game boards are neoprene, not cardboard, and printed poker chips are used as both meeples and tokens. The quality of the components has an obvious effect on price.

With many games, you don't get a choice: The game is the price it is with the components it has. I happened to me that I bought a game at a low price, but then was disappointed at the quality of the components. Great Western Trail: El Paso is a game I like the gameplay of, but dislike the shoddy material. Parts of that game I replaced with my own materials, like cubes and discs; but I can't do anything about the cards and the game board. On crowdfunding board games, the trend is more towards the other direction, I pledged for some games where I ended up with lots of plastic miniatures I didn't really felt necessary.

In the end that is a business decision by the game company: While a prettier game with better table presence can attract some people, others might then be turned off by the high price. So increasingly crowdfunding projects have several options, like a base version with cardboard standees and a more expensive version with miniatures. And of course there is usually a completely over the top "all in" pledge with all the expansions, and all the deluxified materials you can possibly think of.

Normally I am more of a base pledge guy. I actually prefer standees over miniatures. I also prefer wooden tokens over plastic tokens. But I also do like wooden tokens over cardboard tokens, and I especially like metal coins over cardboard coins. That is how I ended up with the "Glory of Rome" pledge level for Concordia, because it has wooden meeples and acrylic tokens instead of cardboard, as well as metal coins.

Concordia is the third "Special Edition" of a classic board game that Awaken Realms produces. They made special editions for Puerto Rico and Castles of Burgundy before. With the deluxification being a more recent trend, the original games were all cardboard. And for some games that is enough. I bought a copy of the old Castles of Burgundy for €19, while the special edition costs €90 plus shipping plus VAT, so around €130 for just the most basic version of the special edition. I didn't buy any edition of Puerto Rico.

Not only did I like the game Concordia more than I liked Puerto Rico or Castles of Burgundy. Concordia is also a game that already has a lot of expansions, and if you have everything you'll sit on quite a lot of boxes. The special edition has all the expansions already existing, plus adds a couple of new ones. And they all come in one large box, making storage easier. What I ordered is a luxury item and thereby by definition unnecessary, but I do get a lot of nice stuff for my money.

Ultimately this moves away from a game that one buys to play it, and enters the territory of a collector's item. It is slightly more useful than a Labubu doll, but the gameplay value doesn't really justify all of the purchase price. In my previous post I talked about Arydia, which did cost something similar, but that game we played for 66 hours, so a bit over $3 per hour, which isn't half bad for a board game. I would have to play Concordia a lot to get down to that price per hour. There is some satisfaction to have a game complete with everything (even if I didn't take the €330+ "Opus Magnum" pledge, as I didn't want the plastic miniatures), with deluxified materials making it look good. Not a reasonable purchase, but sometimes one wants to spoil oneself.

I now put the whole Arydia game back into the box. What I didn't do is resetting the game. Arydia is a game which changes while you play it. Map hexes get turned over when exploring them, cards move from active to inactive or banished, additional game components are unveiled during the campaign. We have a big plastic bag full of cardboard tokens representing gear and other stuff we used and discarded during the game. Arydia is a "green legacy" game, so none of the changes are really permanent. It is possible to reset the game and put everything back where it was at the start of the game. The problem is that this would take hours of work, and I am not really motivated to do that.

I think we played through over 90% of the content of Arydia. Playing the game again would follow largely the exact same story, even if the order or individual battles could play differently. A second playthrough would certainly be a lot less interesting than the first playthrough. And as my board game library resembles my Steam library, too many games - too little time, I don't really see me playing Arydia a second time, even if I enjoyed the first campaign.

Unlike a Steam game I played and don't plan to come back to, a board game takes up physical space. Arydia is a huge box weighing 11 kg. I wouldn't even want to sell it used on Ebay or elsewhere, as I don't want to ship such a huge box. And with the game being in a state where it would need hours of work to reset, I doubt there would be much interest for it anyway.

So I am currently leaning towards throwing the game away. Maybe keep some of the materials like the painted miniatures for use elsewhere, and sort the rest into plastic and paper waste. But that feels somewhat ungrateful towards a game we had a lot of fun with. On the other side, I can't have too many games I'll never play again block all of my shelf space.

If you look at the totality of all jobs, you will realize that most people have jobs in which they manipulate physical objects in some form as core function of their job. Take the stereotypical entry level / student job of flipping burgers. ChatGPT can't flip a burger. You would need an AI-powered robot to flip that burger. And between the cost for the robot and the cost for the AI, it isn't obvious that the minimum wage human doing the job isn't actually a lot cheaper than the robot. Also, even an entry-level human employee tends to do better when things go wrong. Imagine the flat-top griddle having an electrical failure and not heating anymore. A human would notice instantly that his burgers aren't sizzling anymore and react in some way. A robot might happily keep flipping raw burgers, with catastrophic consequences when that raw burger reaches the customer.

In the absence of physical object manipulation requiring a robot, from the remaining jobs there is still a very large part that requires direct interaction with humans, typically customers. You don't need a robot for that, but you need your AI to not mess up by hallucinating. There are famous examples, like the Air Canada chatbot inventing a refund policy, and the airline being forced by the courts to pay that refund. Basically the courts found that a company is liable for their employees messing up, regardless of whether that employee is human or AI. That can get expensive for a company, when a tech-savvy customer persuades your chatbot to sell him a $70,000 car for $1, an error a human salesman is unlikely to make.

If I use the version of Google Gemini that is included in a regular Google search, or I use a regular ChatGPT chatbot, that use of AI is free. That is the typical tech business model of reaching a dominant market position by providing a service at a loss. Which is inevitably followed by some form of "enshittification", because sooner or later the company needs to provide its services at a profit to themselves. Those trillion dollar AI data center investments don't pay for themselves if you don't charge your customers sufficiently. AI companies already charge "pro" users and commercial users, and some companies had to put caps on AI use, because AI use quickly got too expensive. So even for the jobs that an AI can do well, the question is how much cheaper it is going to be than a human, and how the quality and reliability of the output compares in the end.

Sam Altman is predicting the Billion-Dollar Company of One, a startup built by one person with a laptop, an internet connection, and an army of AI agents, worth a billion dollars. While I don't think that is impossible, it is obvious that there are huge parts of the real economy in which such a company is unthinkable. The company can't be involved in any manufacturing, can't be involved in any handling or delivering of physical objects, and can't be involved with any services to humans that can't be done purely online. Robo bricklayers? Robo hairdressers? Robo plumbers? All of these are still decades away even in the most AI-positive scenario, and might never become a practical reality out of cost reasons. I am absolutely positive that AI will not take *all* of our jobs. Although we might end up with some surprises on which jobs AI eventually will replace; for example I can envision a future in which much of porn is AI generated, putting a lot of pornstars out of work. But if you look at a general classification of occupations, you immediately see a huge number of jobs that AI is extremely unlikely to perform in the foreseeable future.