2026-07-31 08:20 UTC
Replies (16)
-
@GustafIII@mastodon.nu 2026-08-02 08:33
@bert_hubert@eupolicy.social This was a very good article I think. There's just one thing I think you missed. The enthusiastic person who comes with their new app or whatever, hasn't created something they couldn't create before. AI has created something for them, that they still can't create. And they probably know this, because they get really mad at you when you point it out.
-
@4dtommie@mograph.social 2026-07-31 08:44
Claudeviolin here. Great piece. I agree: the blind AI rollout is management theater. But then, what isn’t in large corporations? Amusing is the sudden Big Tech conscience. Meta, Netflix, cloud storage, 30,000 photos online: all fine, because we enjoy them. Use Claude to do two days’ work in two hours: suddenly immoral. Energy and copyright are real issues. But judge AI like industry, transport and streaming: what does it cost, deliver and replace and who benefits? Not: new + power = bad.
-
@MarcusEnger@hessen.social 2026-07-31 08:47
@bert_hubert@eupolicy.social what a great article, thank you for a) creating it and b) sharing it in English. I would have missed it without the translation. I gained a lot of insights from it and will strengthen my insistence of defining the (measurable) goal of our AI ideas/prototypes.
-
@fazalmajid@social.vivaldi.net 2026-07-31 09:05
@bert_hubert@eupolicy.social "most LLMs are built on the exploitation of underpaid low-wage workers" Yes, but for data labelers in, say, Kenya, it's still preferable to back-breaking farm labor. The root cause of the exploitation is frankly racist immigration laws that don't allow them to move to better economic opportunities and maximize their human potential. Sadly the political winds throughout the developed world are blowing in the wrong direction.
-
@fazalmajid@social.vivaldi.net 2026-07-31 09:05
@bert_hubert@eupolicy.social The intellectual property argument against AI is the one I find most puzzling. We were on the losing side of the copyfight with publishers, music labels and film studios passing draconian laws like DMCA trampling over fair use, or in France extorting a tax on all storage so a portion of the cost of your MRI goes to Taylor Swift (or Carla Bruni). They have also extended copyright terms beyond all reasonable justification. They are all marginal businesses that barely register in economic statistics, but with disproportionate cultural influence so the tail wags the dog, and we seem to have internalized their specious argument that copyright infringement is tantamount to theft. Here in England, they have even molded the school CS curriculum to indoctrinate children early. Fortunately US courts have consistently ruled training LLMs on copyrighted content is sufficiently transformative to be fair use, and I applaud them putting the copyright lobby back in its place.
-
@david_chisnall@infosec.exchange 2026-07-31 11:07
@bert_hubert@eupolicy.social I stopped when I got to the second sentence because you're doing the same thing AI boosters do to justify the marketing: not defining your terms. What does 'AI' mean, for the purpose of this article? If you include all systems that use machine learning as 'AI', then 'AI' has added a huge amount of value to the economy, most of it prior to the current bubble. If you count only the things that the major vendors are pushing, then you have a three trillion dollar hole. You cannot meaningfully talk about 'AI' while veering madly between these two extremes. It's like writing an article on digital transformation that justifies that game consoles contribute to the economy by taking a segue to point out how industrial automation makes factories more efficient. They both use chips that run algorithms, but that's about the only commonality. The economic impacts of the two (both positive and negative) are completely independent. The only reasons to conflate them are that you don't know what you're talking about or that you're actively trying to sell something by misleading your audience.
-
@CheetahFluff@nyan.network 2026-07-31 12:02
@bert_hubert@eupolicy.social this is pretty much where I've been getting over time. There isn't a way for me to ethically use big models right now, and I won't be very inclined to use it when there is an ethical option because I probably won't have a good use case. But I won't begrudge others for their feelings on using it. The tech has a wow factor and I get that; I might ask them to question the efficacy, accuracy, and ethics though. This was a good read. I can solidify my thoughts better with it, and it'll be a useful thing to point others towards.
-
@ohyran@social.piewpiew.se 2026-07-31 12:47
@bert_hubert@eupolicy.social "LinkedIn thought leadership" and a new level of darkness opened up in the gaping pit that's reality
-
@bartolialberto@infosec.exchange 2026-07-31 14:00
@bert_hubert@eupolicy.social Great analysis, thanks. What I liked most: ""Until now, writing those documents has been an integral part of an organization’s thinking process. Formulating them made both the presentations and the plans themselves better. Even if a document wasn’t widely read, the fact that someone had written a substantial piece about an idea was a strong indication that they had genuinely invested time and thought into it. With AI, however, it’s possible to produce superficially convincing documents with very little effort or reflection Such a document is no longer evidence that someone has thought things through or put in significant work." https://berthub.eu/articles/posts/ai-for-decision-makers/#:~:text=Until%20now%2C%20writing,effort%20or%20reflection"
-
@Di4na@hachyderm.io 2026-08-01 10:45
@bert_hubert@eupolicy.social the one thing I would add is that I think there are better predictions than the one you offered as example at the top. But it doesn't really matters for you point and the rest of the post so eh. Who cares. Thanks for that. I am less positive than you are, for multiple reasons, but I think the end result is more or less the same.
-
@AkaSci@fosstodon.org 2026-08-01 13:14
@bert_hubert@eupolicy.social My use of AI models is weak compared to others. But I do review docs, analysis and code from others, generated with AI assist. My "feelings" so far - AI seems to do better than the mediocre employees, who seem to be totally in awe. My reaction invariably is that quality of the doc or code is poor. The employees spend little time trying to fix or improve it. Writing and thinking skills aren't improving. We need to provide better training on how to use AI effectively. ☹️
-
@ewen@social.ewenbell.com 2026-08-01 14:14
I really enjoyed that read Bert. I appreciate that using a computer to translate a large chunk of text is practical and far from controversial. Whether that's an AI model or another translation method, matters little to me. I used to work in medical research, and the ability to extract signal from noise is the foundation of many scientific endeavours. Some computer tools were useful for this, but AI systems are more suited to the task. Using AI to predict which epitope is more likely to be antigenic is helping laboratories develop ebola vaccines in record time. But this usage is not what the industry is focused on. When they talk of AI's future they are thinking Generative AI and replacing people jobs. I don't they are honest about how much they have sold the technology on reducing company payrolls. So there is a reality check to be paid sometime soon perhaps. I stopped watching or reading the news or other media over a decade ago. It was driven by clickbait and not good for my mental health. So I have to laugh at many of the media claims regarding AI. It's annoying, because there are legitimate uses for such technology, but not so much for the bubble.
-
@cosi@mastodon.sdf.org 2026-08-01 15:08
@bert_hubert@eupolicy.social > Someone who has never been able to program returns to the office on Monday with a fairly functional app. (...) It would be like me swallowing three pills and suddenly speaking fluent French. No, that would be like you swallowing three pills and suddenly having someone speak fluent French for you ;-) (Brilliant article btw.!)
-
@Logical_Error@fosstodon.org 2026-08-01 22:56
@bert_hubert@eupolicy.social a big thing that this article misses is local ai chinese models have largely caught up to the american proprietary frontier models and address a lot of these concerns. and even if the companies disappear, the models themselves will always be available. train once, run and use forever on ur own computer ive written about my opinion and usage of ai here if ur interested: https://logicalerzor.codeberg.page/blog/ai-kernel-development/
-
@Petr90@mamutovo.cz 2026-08-02 04:48
@bert_hubert@eupolicy.social 👍
-
@draconacht@infosec.exchange 2026-08-04 06:21
@bert_hubert@eupolicy.social putting on my nuance hat