@toolbear

The IMF publicly warning about the dangers of the bubble.

Overexposed tech companies collectively losing $1 T in market cap in under a week.

@dilmandila

I don’t imagine using these, but I’m happy with a scenario where Microsoft’s desire to maintain its dominance in the office-suite market and its desire to sell everyone bullshit generators provide directly contradictory incentives.

@futurebird Huh, two books recommended in this thread so far. Earlier today, I bought one and the third in the series of which the other is the first.

@ryanc Given some of the newer tempest-style attacks, it might have been quite accurate naming.

@skavish @mastodonmigration

At this point, the cognitive decline cause by use of LLMs is also well documented and reproduced by multiple studies, so I think the analogy holds quite well.

@tasharoni

I wonder if there’s a market for an ‘EU export quality’ label in USA food: this food complies with the safety standards required for export to the EU.

@JohanEmpa

I don't think the places I wrote about them 20 years ago still exist, but I can.

@gmacon

No, you've correctly spotted why laws are harder to write than toots. That said, I think the example here actually doesn't matter because, according to the CRA, if you want to ship an open-source library then you are responsible for that code.

But the intended wording was third-party services that would count as sharing data with other entities under the GDPR.

@jbqueru @EUCommission

Great example of why this is necessary. Transport infrastructure would definitely be in the 'essential services' category for something like this.

@EUCommission

The Digital Markets Act is a nice idea but it's ignoring how these platforms became so powerful. The problem is that other big entities use them as their sole (or primary) means of communication.

How about defining a set of rules that apply to:

• All governments
• All companies over a certain EU turnover.
• All companies providing essential services (e.g. banking, utilities).

Where the rules are, by the end of 2027:

• They may not require customers to create an account or share data with a third party to communicate with them.
• If they distribute an application for any platform, they must do so via two independent channels.
• Applications must not depend on anything other than the core platform.

Where 'independent' means: does not depend on any approvals from the same vendor.

This would mean that no big organisation in these places would be able to ship apps via the Apple App Store. None of them would be able to ship Android apps that depended on Google Play Services and none of them would be able to ship Android apps at all with Google's proposed lock down mechanism.

And then, suddenly, you have compliance departments in a bunch of companies with deep pockets lobbying Google and Apple to provide an alternative mechanism for distribution.

Note: These are exactly the same as the rules I proposed 20 years ago. If they had been enacted back then, I suspect the DMA would not have been necessary.

@madengineering @lethe

I can’t promise it never happens, but friends who consume such things have assured me it’s a myth. They aren’t short of customers. And, if they gave out a free first hit then they’d need to track who they’d given samples to, which is generally not something black markets of any form want. Additionally, if each drug dealer gave out free samples, there would be a thriving secondary market of people going around towns getting free samples from all of the dealers and selling them on.

@lethe

This is actually misleading. No one gives away free cocaine to get people addicted.

RE: https://toot.cafe/@baldur/117364953454550116

This should be read by everyone participating in the #FreeBSD LLM policy working group.

@billiglarper @mikaeleiman @jon

And yes, that's 82 dead birds in 80 turbine years.

How does that compare with large glass office buildings?

@hanscees Finding instructions for making bombs and other weapons has been trivial for pretty much my entire life. The difference with LLMs is that they make it easier to confuse information and misinformation.

@drwetter @liw @tante

They also make it harder to spot real information.

Go to a library and there’s a pretty good chance that chemistry books are reviewed by actual chemists and will contain correct instructions.

Go to the Internet and there will be a bunch of things like The Anarchist’s Cookbook full of recipes that are likely to blow your hands off if you try them, but also university sites that contain vetted educational information.

Go to an LLM and the part of the latent space that generates your answer might come from a textbook, might come from some novel, might come from some TimeCube-like alternative chemistry craziness. The second time you ask might come from a different mix of these. And you have no way of knowing which.

Note: if you look up how to build a nuclear bomb on the Internet, the instructions that you’re most likely to find are a CIA ‘leak’ that was intended to send the Soviet nuclear programme down a blind alley, so also won’t work. I wouldn’t be surprised if these ended up with many copies in LLM training sets.

I don’t understand the panic over ‘LLMs can tell people how to build biological weapons and plan assassinations’. So can Tom Clancy novels. And both sources will tell you things that are more likely to get you killed than achieve your nefarious goal.

@beeoproblem @glyph

I encountered Rails after WebObjects and my impression was that it was a slow reimplementation of WebObjects with a much worse developer experience. Steve Jobs’ keynote where he introduced WebObjects and the Enterprise Object Framework had a demo that was basically the same as the first learn Rails example and it required about half as much code, eight years before Rails was released.

But everything I read from the Rails people talked as if they’d invented all of these ideas and not just made a slow and bad clone of an existing product.

I largely ignored Ruby and Rails as a result, my only experience with the rest of the adjacent Ruby ecosystem was that they kept claiming to have invented things that Smalltalk had been doing for 20 years before the Ruby people encountered the ideas, in a language whose implementation somehow managed to be slower than the slowest Smalltalk implementation.

@amydiehl Do you know how this translates outside of the trades? Since gender pay-gap reporting became mandatory, the pay difference in computing-related fields has mostly become a promotional-velocity difference: Women are paid at the same rate as men within a level, but it takes them longer to move between levels (or leave because they’re fed up of less-competent men being promoted past them). The net effect is the same but now companies can claim to have a gender pay gap of under 1%. I presume unions would have access to company-wide level and time-in-level information and so would be better placed to call bullshit on this kind of behaviour.

@drahardja A lot of solutions I’ve seen proposed to complex problems have misunderstood the root cause of the problem, economics, psychology, or (in some cases) all three. As a result, the best that they will accomplish is adding complexity to a system but not actually addressing the problem (which may make it harder to actually fix). At worst, they will exacerbate and entrench the problem.

Doing nothing can be better than doing something if the something is counterproductive. If you have an oil fire on your cooker and someone is about to pour water on it, stopping them won’t put out the fire but it’s still the right choice.