I'm not sure this is very easy to show this is a breach of non-discrimination requirements, like under Council Directive 2000/78/EC for employment.
Due to acting like an irrational gambling machine, I agree it can have unwanted indirect discrimination effect in general. But it will probably not differentiate "on the grounds of religion or belief, disability, age or sexual orientation". It is possible, but that would take a lot of work for the lawyers to prove to the court.
I believe the more interesting part is that the EU AI Act (still not in force in this regard until 2 December 2027). This will be clearly a high-risk AI system: "AI systems intended to be used for the recruitment or selection of natural persons, in particular to place targeted job advertisements, to analyse and filter job applications, and to evaluate candidates".
Which does not mean prohibited, but it could later turn out that LLMs will be excluded from being used in high-risk AI use cases (falling under article 6 with no exemptions).
Considering that none of the standards are published yet, I have absolultely no idea how they will ensure compliance with the following parts of Article 10 when using LLMs for such tasks:
"(f) examination in view of possible biases that are likely to affect the health and safety of persons, have a negative impact on fundamental rights or lead to discrimination prohibited under Union law, especially where data outputs influence inputs for future operations;
(g) appropriate measures to detect, prevent and mitigate possible biases identified according to point (f)"
I don't think that's technically possible to do so with LLMs in general at the moment, even with the full cooperation of the model providers. Maybe you can do some meaningful audits for smaller models. But the EU AI Act may end up excluding all the generic "using-LLM-but-not-entirely-sure-why" vibe coded approaches from high-risk use cases (in Annex III). Which would make sense.
So, for many countries, it's not just that sealed documents are not accessible, many people in traditional democracies couldn't imagine and bear this level of publicity for every court document like you have in the US.
The difference is quite subtle in common law v. continental law systems.
In continental traditions, nobody really says a higher court decision is binding in the same sense as they do in common law. But even in common law, lower courts have the power to distinguish a case from a binding decision from a higher court. Or later courts may state that a particular argument used by a judge in the higher court was just 'obiter dicta' and not part of the 'ratio decidendi'. In continental systems, for most court processes, higher courts do have a strong persuasive power without formal binding (sometimes it may be stronger than your average "obiter dicta" in common law). So, there is a whole spectrum in obiter dicta as well.
This decision was a preliminary injunction only from Landgericht Munich - you will have a final injunction order:
https://the-decoder.de/wp-content/uploads/2026/06/26_O_869_2...
In Germany, you also have the Oberlandesgericht and the Bundesgerichtshof as a higher venue for civil cases (and the Constitutional Court as well).
So, this is not like "the law has changed" at all.
(I'm not a German qualified lawyer, but qualified in a close enough continental system, and on my way to be qualified in a common law jurisdiction soon, that's why I dare to comment on this aspect)
Before suing all your coworkers for garnishing their wages, I suggest you first consider how much fun you will have when you have to pay for the legal costs of the other party. But you will find that out for yourself.
Anyways, I don't think it's your choice whether you want to take on huge law firms - that's usually the other party's choice.
(I do hope most security researchers have a bit more training than "some education".)
If you involve a lawyer only for the final review, they will understand you don't want them to rethink your approach nor to advise a different strategy. So, they will just bill you for the time, which seems to be fine for both parties of the trade.
And, where it's mandatory to have a law firm draft a deed or a contract (like for mortgage or real estate transactions, depending on the state you're in), your Claude-made first draft will not make your bill lower.
Just don't try to rely on this for adversarial work, like in litigation.
BTW, just curious about the timing: how could Claude for Legal save you thousands of dollars at your last company if the product came out in February? (Not that Claude for Legal is any special in its legal output compared to "ordinary" Claude)
I don't understand how different this startup is from established, non-John Deere segments, like MTZ?
Definitely low tech and around the same price as the startup:
https://www.mtzequipment.com/products/tractors
The Council of Bars and Law Societies of Europe (CCBE) has adopted a technical guide for lawyers on the use of GenAI.
It is addressed to lawyers and bars/law societies (which, in most countries, also regulate lawyers).
It is a clear and deliberate step to educate legal professionals on more technical matters that were previously rarely a focus of their training.
The purpose of the paper is to better acquaint lawyers with their options, including the possibility of using local models and understanding the trade-offs.
CCBE was created by the bars/law societies as a representative body of the profession in Brussels, to whom the EU institutions can turn.
I believe this is an important step in toward better dissemination of the use of AI models.
What makes it notable is the level of technical detail that will hopefully help lawyers formulate what they need - comparing the options of on-premises, colocation, IaaS or fully managed/API), and giving indicative hardware budgets (dated September 2025) ranging from ~€2,000 to €20,000 for local inference boxes. It also discusses quantisation (FP16 vs INT8 vs INT4), context-length trade-offs, and tokens-per-second thresholds for interactive use.
Curious to hear your views on whether there is any point in going in this direction. Or if you believe anything important is missing or incorrect.
They both work in the same market but they have pretty different careers and understandings.
I simply can't believe why on Earth would people choose Altman over Amodei to trust in these kind of pretty important questions.
This is not about who is the more savvy investor maximizing shareholder value. I personally don't care whose company grows bigger or goes bust first, OpenAI or Anthropic.
The real stakes are different, and Amodei is better suited to be trusted in his decision.
Unfortunately, the best choices do not seem to fit well with either the federal political climate or the mainstream business ethics in Silicon Valley.
Not that our opinion would matter...
In a movie, I'd definitely involve Ötzi as well (https://en.wikipedia.org/wiki/%C3%96tzi). Ötzi was found like 30 km from the impact site. And could have been a contemporary.
E.g., he cursed the guy who shot him and whose village is struck by a meteor in the end.
I don't understand: if there is a shortage of doctors, why are we trying to solve that by training AI models modelled on influencers that spits out (hopefully improving) advice at 10x the rate of a human doctor? Is it impossible for highly advanced societies like ours to pay more for people to get trained as doctors, nurses or whatever is missing? Or to convince them to choose a profession that deals with other humans instead of UBI?
I don't think people are afraid of doctors using imperfect tools. That is the easier part. But that will not solve the problem of too many patients for a single doctor and what leads to the lack of empathy. This was a problem even before AI. It seems society does not have empathy for these kind of "professional problems". Offering tools instead of humans is an even riskier approach, not for that particular individual, but for how society tends to build trust and empathy.
We tend to see everything now as a problem with a technical solution because we only have confidence in solving technical problems.
Losing brand control (IP) and lack of control over your sales channels is a big problem for traders who want to comply with dozens of regulations, exclusions and restrictions required by manufacturers, other sellers, trademark owners etc. The article also mentions means tax and customs compliance problems, but this also affects other trade law related issues, like competition law problems (antitrust). It is surprisingly easy to breach these even for tiny sellers. You don't want exposure to these kinds of sales. It is never just selling, it is always also complying with lots of new laws, And rest assured, Amazon bots will not work that out for you.
And now, these guys have to scrutinize the email orders as for their source because all the compliance process they have built in their webshops or wherever they advertised, will simply be bypassed by Amazon bots.
Terrible idea.
It's not about invalidating your conclusion, but I'm not so sure about law having a right answer. At a very basic level, like hypothetical conduct used in basic legal training matrerials or MCQs, or in criminal/civil code based situations in well-abstracting Roman law-based jurisdictions, definitely.
But the actual work, at least for most lawyers is to build on many layers of such abstractions to support your/client's viwepoint.
And that level is already about persuasion of other people, not having the "right" legal argument or applying the most correct case found. And this part is not documented well, approaches changes a lot, even if law remains the same.
Think of family law or law of succession - does not change much over centuries but every day, worldwide, millions of people spend huge amounts of money and energy on finding novel ways to turn those same paragraphs to their advantage and put their "loved" ones and relatives in a worse position.
Zucman taxes rich individuals (100m€+), not Mistral. AI Act rules are not that difficult to comply with by GPAI model providers as long as the model doesn't become systemic risk... They have to spend a lot more time on PR and handshaking with French politicians than on AI compliance. They probably don't even have a single FTE for that... So that's just prejudice I believe.
Socrates said no such thing, no writing of Socrates has survived. He was just a character is Plato's book, Phaedrus.
Please do find the original paragraphs before accusing Socrates of this. https://standardebooks.org/ebooks/plato/dialogues/benjamin-j...
Of course, you can read and interpret that same book a thousand different ways, like he was talking about knowledge not being the same as writing things down, or whatever you want. But we don't even pretend to read the things we talk about. We just repeat nice narratives we have supposedly read somewhere else, digested by someone else, somehow.
Digital Spaceport is a really good channel, I second that - the author is not sparing any detail.
The cheaper options always use CPU only, or sharding between different cheap GPUs (without SLI/switching) - which is not good for all use cases (he also highlights this).
But some his prices are one-off bargains for used stuff. And RAM prices doubled this year, so you won't buy 2x256 GB DDR4 for $336, no matter what: https://digitalspaceport.com/500-deepseek-r1-671b-local-ai-s...
Yes, there seems to be lots of mistakes and no easy way to mark it. Highly endangered: Malayalam (=35 million speakers), Hungarian (14 million), Uighur (11 million), or Swedish as endangered... These are quite obvious mistakes even for a layperson.
I agree that this is a very exciting and really crucial research and I'm glad there is funding for this.
But it's very strange that Hungarian is marked as "highly endangered" at https://aidemos.atmeta.com/omnilingualasr/language-globe
Highly endangered is supposed to mean "The language is used by grandparents and older generations; while the parent generation may still understand the language, they typically do not speak it to children or among themselves."
Then why is Hungarian marked as such? Obviously not true with 14 million active speakers and being the 20th in terms of the most language resources published on the Internet.
Additionally, the feedback mechanism seems also broken ("There was an error submitting your feedback. Please try again.")
That still seems to be a problem.
It was not what "Socrates thought", but what Plato put into Socrates' mouth in Phaedrus, and even this imaginary Socrates is not saying anything like that, just referencing an even earlier Egyptian tale: "There is an old Egyptian tale of Theuth, the inventor of writing, showing his invention to the god Thamus, who told him that he would only spoil men’s memories and take away their understandings..." https://standardebooks.org/ebooks/plato/dialogues/benjamin-j...
But that's just pedantery. The real painpoint is that just because there are lots of useful AI tools, it doesn't mean it's not dangerous at the same time for a surprising number of 8B people currently alive (children, elderly, mentally lazy or just fatigued).
At the very least, they will end up being exploited by bandits. And if you let the bandits continue to exploit those who lack certain mental resistance, the bandits will become stronger etc.
Due to acting like an irrational gambling machine, I agree it can have unwanted indirect discrimination effect in general. But it will probably not differentiate "on the grounds of religion or belief, disability, age or sexual orientation". It is possible, but that would take a lot of work for the lawyers to prove to the court.
I believe the more interesting part is that the EU AI Act (still not in force in this regard until 2 December 2027). This will be clearly a high-risk AI system: "AI systems intended to be used for the recruitment or selection of natural persons, in particular to place targeted job advertisements, to analyse and filter job applications, and to evaluate candidates".
Which does not mean prohibited, but it could later turn out that LLMs will be excluded from being used in high-risk AI use cases (falling under article 6 with no exemptions).
Considering that none of the standards are published yet, I have absolultely no idea how they will ensure compliance with the following parts of Article 10 when using LLMs for such tasks: "(f) examination in view of possible biases that are likely to affect the health and safety of persons, have a negative impact on fundamental rights or lead to discrimination prohibited under Union law, especially where data outputs influence inputs for future operations; (g) appropriate measures to detect, prevent and mitigate possible biases identified according to point (f)"
I don't think that's technically possible to do so with LLMs in general at the moment, even with the full cooperation of the model providers. Maybe you can do some meaningful audits for smaller models. But the EU AI Act may end up excluding all the generic "using-LLM-but-not-entirely-sure-why" vibe coded approaches from high-risk use cases (in Annex III). Which would make sense.
https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng