I find the idea that people can use LLMs to exercise their rights as citizens appealing however. Many people aren't aware of the rights they have, and LLMs are pretty good at surfacing some stuff without having to pay lawyers. Having to hire a lawyer is imo actually a huge way of gatekeeping people from exercising their rights. I heard a lot of local German public institutions are currently being flooded with people arguing their case with the help of LLM that they previously weren't really realistically able to do. So I don't see it all as bad.
>are pretty good at surfacing some stuff without having to pay lawyers
They are also really good at making stuff up as evidenced by the many, many, many examples you read in the news about actual lawyers using AI to write briefs that are full of errors and hallucinations.
In civil courts, you'd likely get more sympathy from a judge if you represented yourself and admitted your lack of understanding, rather than try to appear as someone you're not because you wrote some prompts and copied the output.
Current court systems around the world are just not built to handle the flooding of cases from the citizens.
The stupidest analogy is open source projects having a hard time accepting LLM generated PRs from the masses, because review process is the bottleneck.
No idea how to fix this, to be honest. In coding world, with some mental gymnastics, I can see code not being reviewed by people anymore. In courts, things generally have more consequences, and you can’t really roll back decisions that easily.
you could probably write a cool little gotcha of an SF short story about a barren wasteland of a planet that keeps broadcasting out legalese that's revealed to just be LLM chatbot lawyers pedantically arguing with one another about xeno legal doctrine
The legal system, which is a machine/technology by itself, will be eaten out. I wonder what will replace it. Botnet law arbitrage? – Personal assistants constantly negotiating with each other to avoid permanent civil lawsuits?
the cost of making a legal argument can collapse while the cost of reaching an enforceable, legitimate decision may go higher, which will gate the "justice" system even more.
It's really not. The legal system is slow and inefficient because it's a deeply pipelined system built to maximize the throughput of the bottleneck resource: judges. Judges are constitutional officers who exercise independent authority in meat space and thus are necessarily limited in number. The rest of the design flows from that.
If you got the judge, all the parties, and all the witnesses in a conference room together until the case was resolved, you could probably handle a lawsuit in a few months. But each judge has hundreds of cases pending before them, so that would never work. Instead, you get something like how a GPU works. You do some work on a case, submit the work to the court, then work on something else for a few months while you wait around to get the results back. Then you do some more work and submit it to the court, then go do something else for a few months while you wait to get the results back. A few months of actual work gets spread out over a few years that way.
Why stop there? The bottleneck is not some immutable force of nature. The number of judges is determined by legislative action. It is within Congress's power to allocate new federal judicial seats, and likewise at the state level with the equivalent lawmaking body (for most states, at least). Why don't they do so?
Because federal district court judges have a tremendous amount of power. You both don't want 10,000 of them running around and you don't want to water down the qualification for the position--if you did do that, then that would devalue their judgments, everyone would appeal, and you'd just shift the bottleneck up to the appellate courts.
At the end of the day, litigation is conflict resolution. You don't just need a decision, you need a decision from someone authoritative enough to bully Fortune 500 CEOs into submission to accept the judgment.
You presented the concern from my adjacent comment perfectly (“LLM performance: common law vs civil law” essentially). So is AI possibly just growing the “Reverence for Professional Experience” factor that plays such a big role for legal compensation here in the US?
Creating virtual products get cheaper, so shouldn't it even decrease if everyone vibe codes their app for less than a dollar instead of hiring a dev team for 50k or ordering a white label app or spending 3 dollars to buy an existing app?
I don’t know in the USA but in France, if it’s deemed that you launched a lawsuit knowing very well it wouldn’t succeed, you are susceptible to get a 10k€ fine. Even jail in serious cases.
Yes. And you(r lawyer) can collect lawyer's fees and you can be made to pay the court fees, if you lose.
IME the American legal system is set-up to discourage litigation, though. A common tactic is to bury your opponent in the threat of heavy damages or jail-time to get them to settle for what you were originally after, which courts are perfectly happy to facilitate because it gets a potentially lengthy trial off their dockets. They'll punish (or be biased against) whichever party seems responsible for not accepting a "reasonable" settlement.
Algorithmic abuse of the system to extract payments already exists in the form of the debt collection industry.
Why the hell are they not fighting fire with fire? This is not sustainable. But it is annoying that the AI labs get to play arms dealer, selling to both sides.
They don't have a choice. If you're a firm and you work with IE a bank, the bank uses AI, their whole supply chain uses AI- your counterparty uses AI- it's a classic arms race with even more social pressure involved (since daily news keeps banging on about AI).
In case it's helpful to anyone, MiSTer is a project[0] that recreates classic computers and video game consoles on hardware.
What do I mean by hardware?
Rather than emulating the device in software, which can cause various issues (especially related to timing), by mapping the actual device logic to a field programmable gate array (FPGA), you can achieve an "exact" replica of the hardware.
It's a super cool project, and I highly recommend checking it out if you have an interest in old machines (Apple II, Commodore 64, etc.).
"exact" is still in quotation marks because things like open bus behavior still have to be explicitly simulated. It's still a very accurate simulator, not an exact replica - much more accurate than most fast software simulators can be, though.
Worth noting that timing-exact emulation is for, more or less, less than 32 bit consoles. The timing of newer hardware is so unpredictable that nothing can reasonably rely on details, which is just as well: it would probably not be possible to emulate it precisely at an acceptable speed.
PCs have been "unpredictable" like that almost since the beginning due to hardware variation, so emulating timings isn't much of a problem for emulators, as is the case in the featured article.
Check out the Superstation One [0]. Outer shell modelled after the original PS1. Optional optical disk dock. Simplified UI. Lots of outputs. Well priced.
Super cool! To be honest I'm not into emulation/retrogaming so much (even though I do have a Pimoroni Picade) but the fact that this exist as a production is very inspiring, thanks for sharing.
If you want something more powerful than MiSTer and have a bit of patience, there is the Replay R2 being developmed around the Agilex 5.
- Agilex 5 E-Series 043 (Replay2 phat) has 434,240 logic elements (LEs).
- Agilex 5 E-Series 013 (Replay2 lite) has 138,060 Logic Elements (LEs).
The website (https://www.fpgaarcade.com/news-2/) hasn't been updated in some time but the dev now has prototype boards that he's in the middle of bringing up. It is purpose built for arcade emulation and has tons of I/O, lots of memory (both SRAM and HyperRAM), and (on the phat) a PCI connector. One or both (can't remember which) accepts an ARM daughterboard for hybrid emulation. There's a Discord for it you can track down to follow its progress basically in real time as mikej progresses.
There are much cheaper clones of the DE10 available these days. For a little while fully assembled MiSTer's were available on AliExpress for around $150-$200 although it looks like maybe they dried up. There's also the SuperStation based on the MiSTer project for around $250 if you're patient.
Does mister work with kv260 boards "out of the box" or is the official mister target a de10 nano? (I love my mister, assembled oneba few years ago). I don't follow the project closely due to zero free time these days
You have to base your build around either the DE10-Nano or one of the clone boards from QMTech. If I were getting into MiSTer I would skip faffing around with dev board builds and just get a SuperStation One, and I say this as someone who has two MiSTer builds based on DE-10 boards. In my case I'm using a MiniITX motherboard that the DE-10 plugs into called the IronClad Plus. The SuperStation one is apparently awesome. Great IO options, reference quality analog video, and nice industrial design.
Both require having an understanding of what you are trying to emulate. There are systems which don't have an accurate specification so they are not implemented bug free.
LLMs are basically multi-dimensional magic mirrors.
Depending on where you point them, they can be incredibly useful.
They can even be useful when you point them at each other (though increasingly difficult to get good results).
I'm excited for the promise of RSI and a future where models have inherently "live" weights, but it's not clear to me that the transformer is more than a useful tool to help us get there.
I'm also aphantasic and dream normally, and I once got really into lucid dreaming for a ~semester in college.
My best successes would occur after drinking a cup of coffee and immediately taking a nap before a study session.
It was great, I "learned how to fly" and could "morph" my dream into whichever direction I wanted.
One thing stuck out though: whenever I would try and focus on something, I would see a black dot enter the center of my vision. The more I tried to focus, the larger it got. If I focused on an area attentively enough, such that the black circle became all encompassing, I would wake up.
After a few times of this, I could start to notice the focus, and then sort of relax out of it, and thus continue the dream.
It feels like the black dot is my real-life vision coming into focus, and "winning" over the visual aspects of the dream. It's always continuous from black dot to eyes opening.
I rarely lucid dream now, but I've noticed a similar thing happen whenever I try to read text - the black dot appears and I wake up.
Wish I still had unlimited time to go back to sleep!
One of the most horrifying recurring dreams when young was "geometric nightmares". This black circle would expand in size while a deafening silence grows. The volume is a void, and it's deafening.
An impending sense of doom as the expanding black void circle approaches. There's a real fear of oblivion, of ceasing to exist. Then you wake up.
I'm not aphantasic, but that's exactly the experience I get when coming out of a lucid dream. Except it's my expected surroundings not black. Gets a bit weird and inceptiony when it's the wrong surroundings! Given the number of cultural references along these lines, I suspect it's not uncommon.
Focus takes training, just like physical exercise, it is a super power to be harnessed.
You can train yourself up by looking at your hand, then building a body, then building a room, and maintaining all of it. When you are strong enough, you walk outside.
I hit rate limits recently after the Astra launch while using the goal feature, and I was surprised as it was an afternoon of work, something I hadn’t seen with Sol over much longer time horizons.
In this article, I found the supposed lacked of accounting for cached tokens interesting, as well as the wasteful goal/wait behavior, but I haven’t proven either.
It's interesting to me that the agents comparison page[0] doesn't list codex's app-server as an option.
I've found the app-server to be the most flexible, compared to the raw Responses API or Agents SDK.
Certainly seems like everyone is still figuring out the right interface here.
Also of note, since GPT-5.5 or so, Codex doesn't even use the Responses API as intended, but instead a "lite" version where they manage the context more manually (like sending the full transcript or using a custom web.run tool instead of the provided `web_search` tool).
If you follow the docs, it will lead you down a lot of well-intended functionality, but most of it is thrown away in their most successful harness.
reply