As someone who spends way too much time every day trying to stay up to date with developments in the AI industry (next to impossible), here's a couple of (admittedly provocative) thoughts in no particular order to prompt a debate:

  • This is not an accusation, but the vast, vast majority of people are utterly clueless about AI, including those who publicly comment on it in places like reddit. It's people who tried ChatGPT free a couple of times, back when it was all over the news, found it meh and now believe they are qualified to assess the situation. It's people who in 2025 use 4o-mini for knowledge questions, get bad results, don't understand why and then blame the tool for their incompetence. If I had a penny every time I saw the latter! DeepSeek R1 brought some public attention recently, but you'd probably have to ask more than a hundred random people on the street to find one who has actually used it. The rapid progress in AI happens almost entirely under the radar of the public. That's a problem.

  • A rule of thumb: If someone hasn't heard of DeepSeek before January 2025, they're not very involved. If someone can't name the most recent 3-4 models by OpenAI and thinks of Mickiewicz or Shakespeare when they hear "Sonnet", their opinion is as good as that of uncle Wojciech on advances in treatment of COVID-19. Point being: What "most people" think, consensus, is a poor indicator for what's going to happen.

  • For those curious about the current state-of-the-art, here's a demo of OpenAI's 2nd or 3rd best publicly available model, not including o3 which is a few weeks away from release: "Can't even count the number of R's in strawberry".

  • AI is absolutely going to replace a huge number of jobs in the rest of this decade and we should all be concerned regardless of how safe we think we personally might be. It's going to change societies, transform entire industries, impact state budgets, meaning there's no place to hide.

  • Poland will, that's my prediction, get hit sooner and harder than other countries. A good part of the Polish IT sector for example is outsourcing by and for foreign corporations. Those are the people that first get made redundant, not those making hundreds of thousands of dollars a year designing and managing new products and services in the valley. Those with authority to hire and fire will first make themselves unpopular in places far away, before they think about cuts at home, anyone willing to bet?

  • It's far from limited to IT. BPO in general is a massive industry in Poland and many of those jobs are prime candidates to be automated. Giants like Google and Apple (via subcontractors) have large offices in Kraków and other cities, where anyone with basic computer literacy can get the required training in 2-3 weeks and make decent money categorizing brands or labeling map data. McKinsey employs 1400 people here in Poznań providing services to the actual consultants. HSBC Kraków is 6000 jobs for 27 markets in 11 languages according to their website. Banks are among the first to eliminate bullshit jobs, it's happening as I type. TL;DR: BPO alone accounts for about 5% of the Polish GDP. For reference: That's roughly what the car industry is to the German economy.

  • the most foolish take is "AI isn't gonna replace humans". That's true in the literal sense. It's only gonna replace "some" humans. But when work today performed by a team of 20 can be automated with 5 staff for quality control, you're more likely part of the 15 than not. Give it a year or two and once data conclusively shows the particular automation works well enough, the remaining five become two or three. It's not unlikely that one day in the future a balance will be found, but for the short and medium term it's gonna be a massacre. Unlike the steam engine back in the day, AI requires next to no CapEx to use, leading to explosive adaption by businesses the moment it meets requirements. Companies are not going to scale up output (to whom?) at the same rate as they increase productivity per employee. Naivety doesn't buy groceries.

  • almost as idiotic: Various iterations of "AI isn't intelligent", "there is no AI", "AI can't think|reason|create something new", "it's just a bunch of maths". None of that matters. It's all philosophical talk that some might find interesting to debate over a bottle of wine, and there's nothing wrong with that, but these questions have no relevance for the real world. If a computer system can perform a certain function better, faster, cheaper, more reliably or a combination thereof than humans, companyies are going to use it to replace people. Telling your friends "but it's just predicting words!!!1!" is not going to pay your rent I'm afraid.

  • since I'm already here: Poles being "proud" of Polish people building OpenAI is about as smart as Indians being proud that half of silicon valley CEOs are Indian. There's a reason they left and now create immense value elsewhere. That's a tragedy, nothing to celebrate. Out of the six Poles on that famous group photo, all but one left Poland >12 years ago. The first was Łukasz Kaiser in 2004.

  • Europe has nothing, we are entirely dependent on other powers. No, Mistral is no serious competition. "Le Chat" isn't even close to state of the art.. It frankly sucks.

Debate.

by opolsce

14 comments

  3. My clients clearly started using AI for communication, briefing, problem analysis because they sound smart while writing about things that are not applicable for given project or things we already rejected but the new generated message/e-mail do not retain that memory.

    It is fun.

    I think we will start charging consulting fee because it just bottlenecks us so much.

  4. Europe has indeed nothing.
    We debate about plastic bottle caps 😂

  5. Just one interesting point about BPO. Big Tech has been having AI related projects in Poland for a decade and dozens of thousands of people used to work there. It was just never called AI, so nobody really understood the idea behind weird *artificial* teams and departments. Everyone did what they were told to do.

    Still at the moment the main issue of the industry is that Poland became more expensive than international companies expected. 1200-1500$ gross salary in 2019 was pretty good for an entry-level role and people were happy. In 2025 they pay the same, but it is not a comfortable paycheck anymore. So at the moment the trend is to just move teams to India, Romania and other cheaper countries. I wouldn’t be surprised if a low-skilled BPO in Poland dies before AI can even replace it.

    However, I don’t think that Polish IT is in mortal danger as a whole. Companies will still want to pay 70k$ instead of 250k$ to the remaining quality assurance experts or whoever will be needed. For example, my employer stopped hiring for most engineering roles in the headquarters and previous regional hubs and now focus solely on Poland, Mexico, India, Philippines. Poland specifically, because the supply of good engineers is decent here.

    I easily can see a world where there is an American HQ with C-Level and a couple of business people, but all other remaining employees work in some low cost areas. AI will fork absolute majority of countries.

  7. > Those with authority to hire and fire will first make themselves unpopular in places far away, before they think about cuts at home, anyone willing to bet?

    cuts at home make them popular with stockholders, as it can cause stock price rises.

    Cutting at home and outsourcing is nothing new.

    It might not end up being great, but house ownership rate here is high, passport is strong, loads have “alternative” means to keep warm in winters by burning trash, and some alternative means to get free food – people will survive, and it will not cost the state too much to make sure the others survive also.

    I understand exactly what you mean by “bullshit jobs”, but even taking just that out, people invent loads of bullshit dumb activity (toilet paper, instead of water and hand and towel, overpaying for sugary water which is bad for you, and needs to be packed, marketed, transported, stacked etc., stupid/non eco hair dyes/nail paints and so on) and will continue to invent more so

  8. People and jobs will adapt like it happened with many new technologies, back in the day there was other or limited usage for container technology, configuration management, SAS, PAS, agile methods and many things wich make our life easier today.
    This adaptation will take years, when I started in IT I didnt even knew the role I am currently in exists.

  9. > For those curious about the current state-of-the-art, here’s a demo of OpenAI’s 2nd or 3rd best publicly available model, not including o3 which is a few weeks away from release: “Can’t even count the number of R’s in strawberry”.

    That’s down to how the LLMs process their input. They simply are not geared for string manipulation like counting the number of words or letters or reversing text. Arguing that this somehow proves them unsuitable for other uses is a terrible argument in the vein of “AI can’t think|reason|create something new”.

  10. Those companies that reduce junior hiring will be the first ones to overpay the same juniors once seniors leave the company and there is noone to replace them. I „love” capitalism because it bites itself in the ass so much for short-term profits that makes it spend more in the long run.

  11. It certainly will require a drastic changes to economy. If we let money to trickle up to the owners of AI, soon we’ll be left with nothing to buy groceries and pay rent with. It’s also an issue to the capitalists themselves – who consumes when there’s no one with money to spend on consumption? Governments will change too. At this point, if a country doesn’t have potent natural resources, it’s best bet for stable government and rich population is democracy. A person must be skilled and educated to create value. Must be able to think. While a diamond mine can run on starving slaves. Well, a mine and an AI. When humans are no longer the source of wealth their lives lose value. That’s why we need strong mechanisms to stop that. I don’t mean AI – that’s inevitable, I mean the growing divide between the richest top 1% and all of the rest.

  12. Replacing bottom tier call center agents is one thing. But I find it hilarious IT jumped on the whole thing somehow forgetting that you have to be a junior to be a senior… It will be pretty funny when the whole industry will hit a wall of lacking brainpower in positions that actually matter because no-one managed to get through the ladder…

  13. This post is really emotionally charged, you sound like you really want it to be true. Which, I would take with a grain of salt.

  14. Tbf I don’t really think AI is better than humans at the moment, but it doesn’t matter. When a corporate has to choose between hiring someone do make a thing, or get AI to do subpar shit and then hire the same person for half the wage/time as a quality control, it’s going to happen. Already people in journalism are being reduces to that level. We’re going to be drowned in AI mediocrity and get fucked anyway because less money spent on wages = profit for corpos

Comments are closed.