Accountability is the biggest unaddressed challenge for AI implementation.
When one person is able to do too much too quickly, they can create more liability than they can accommodate if something fails.
It is essential that a human is responsible for the utilization of any AI output in the real world, but that is not enough. For our own sakes, we must find ways to minimize the tech-debt bankruptcy blast-radius of those who would utilize (knowingly or unknowingly) AI to create flawed systems upon which others rely.
An example: Jim vibe-codes an extremely popular micropayments app. He hires a few people and sees the company as the WhatsApp of money -- a few engineers and some agentic support staff. It pulls in a few million in VC money -- enough to draw in tens of millions of users. One day, a flaw in the infrastructure causes all of the users' unsalted banking information to be released.
Agentic AI allows that entire list of customers to be exploited rapidly, so the losses for society are in the tens of billions. Jim's company is immediately bankrupt, of course, but there are only a few million dollars to go around.
Today, most of Jim's incentives are to go ahead and build that app. The same is true for his few employees and a small VC contribution. There's not much capital at risk compared with the societal exposure.
How do we ensure that AI users are accountable not just for their actions, but for the size of the risk-exposure that they create?
I have had multiple conversations on HN with people who fight tooth and nail, I mean really ready to die on their hill, because they believe they shouldn’t even have to vet what comes out of an LLM. It’s absolutely baffling to me. The most bizarre excuse is “it codes better than people,” which is not even remotely a given and needs a lot of qualifiers.
I understand there is a push/pull with regards to how much we should let them do, but to not even look at the results before you make them somebody else’s problem? It’s just selfish. There’s no other word for it. You are simply taking the work you were supposed to do it and dumping it on somebody else. These are probably the same people who get upset (rightfully so!) when somebody doesn’t proofread their article/blog before publishing it online.
Everybody wants to use LLM’s to cut corners on their work but nobody wants to be downstream of it. That simply doesn’t work.
>In The Unaccountability Machine, Dan Davies argues that organizations form “accountability sinks,” structures that absorb or obscure the consequences of a decision such that no one can be held directly accountable for it. Here’s an example: a higher up at a hospitality company decides to reduce the size of its cleaning staff, because it improves the numbers on a balance sheet somewhere. Later, you are trying to check into a room, but it’s not ready and the clerk can’t tell you when it will be; they can offer a voucher, but what you need is a room. There’s no one to call to complain, no way to communicate back to that distant leader that they’ve scotched your plans. The accountability is swallowed up into a void, lost forever.[0]
How is that any different from the pre-llm days, when Jim was using stackoverflow to build the largest crypto exchange in the world? Where's stackoverflow accountability?
An exploding job-recruitment offer might not attract the kind of folks we want designing a system that absolutely must work after a decade in space.
I've worked with NASA and ESA employees/contractors who've made technical miracles happen in space. I don't think any of them would be drawn to this style of recruitment.
I got the impression that despite using terms like "mission critical", this isn't about the hardcore technical wizardry behind propulsion and safety.
This is a call for developers of the very long tail of logistics related stuff. I'd imagine a moon base would need someone to write the software for schedulers, dashboards, etc. and engineer the parts that interface with and provide non-critical telemetry to those systems. I'm not saying that stuff isn't hard, but it's not anything life or death.
Some of those roles might not even be technical at all and be more about coordinating the human side of those efforts.
The US administration is, at present, regularly violating the law and ignoring court orders. Indeed, these very releases are patently in violation of multiple federal laws -- they're simultaneously insufficiently-responsive to meet the requirements of the law requiring the release of the files and fall afoul of CSAM laws by being incompletely redacted.
The challenge, as we're all experiencing together, is that the law is not inherently self-enforcing.
> ICE has likely violated more court orders in January 2026 than some federal agencies have violated in their entire existence," Schiltz said, adding that he counted 96 court orders that ICE has violated in 74 cases.
"Allegations" from the exact judges whose orders aren't being enacted? The orders in question are pretty simple: release this guy. Don't take this guy out of state. It's pretty clear when they're not being followed. This guy is not a slouch:
Did you notice that one article I linked involved a DoJ lawyer admitting that she couldn't convince ICE to obey court orders that she was trying to transmit to them? That's beyond an allegation and into admission. How is that not evidence?
The problem is that it's always specific to a particular case. So, if one guy isn't being released according to court order, they could order someone held in the courthouse jail until he is, and probably just the threat will get him released. But then 1) nobody ends up in jail, because they're not in contempt anymore and 2) it doesn't do anything for any other cases, and there are so many other cases. This sort of contempt where a judge can just order it is "civil contempt" and is meant to convince someone to comply with the court order, it can't be used to punish someone longer than that (criminal contempt can, but you need an actual prosecution, trial, etc).
You might think "ok can't they be held in contempt for the pattern of ignoring court orders" and, well, you'd think so. But that looks a lot like a universal injunction or a class action and SCOTUS has deliberately been nerfing those.
If they've simply been committing crimes then judges don't have anything to do- they'd have to be prosecuted by someone, or I guess sued civilly, but that won't put them in jail either and takes forever.
Furthermore, there are numerous allegations that the documents that have been released contain CSAM, which (referencing the PDF above) may fall afoul of 18 U.S.C. 2252–2252A.
In addition, one need only glance at the action in US courts to see egregious violations of the Constitution and valid court orders playing out daily.
Yes, the Abrego Garcia and Öztürk detentions are two very newsworthy cases that have actually reached the point of a final judgement in the district courts, as opposed to "merely" preliminary injunctions against the government.
(It's also worth noting that almost none of the government's appeals to their losses in preliminary injunctions have been on the merits as to whether or not their actions were legal, but rather on the grounds of "no one should be allowed to challenge our actions," which has also been a fairly losing argument for everybody except SCOTUS.)
Allegations are literally evidence. "He attacked me" is an allegation of a crime and is evidence that would be used in conjunction with other evidence to prosecute said crime.
There's more than enough credible reports of CSAM in the Epstein Files dump - more than enough for me to not go and download even a single file of them myself, simply because German law does not care about why you are in the possession of CSAM, even if you took the picture yourself.
The legal situation regarding CSAM is very strict no matter which country, and I better hope no one here will actually be dumb enough to provide actual links.
If those reports are true then what we have is not just an effective deterrent for download and distribution of the set, but legally prosecutable malware targetting anyone who does, empowered by the Interpol CSAM database to which the DOJ should probably already released the offending material.
It's a tricky issue. In many countries it's not illegal and quite common for children to run around naked in public, during the summer on beaches for example, and so millions of people have holiday photos that are technically CSAM in their possession that they don't even know they have.
CSAM has a meaning identical to child porn but doesn't make that meaning explicit. Drawn or generated depictions of child nudity can be considered CSAM in some jurisdictions.
They keep illegally appointing unqualified hacks as US attorney in defiance of the mandate they're approved by the Senate (Essayli, Habba, Halligan, Sarcone, Chattah) - judges have found at least five of the appointments illegal. As one example: https://www.politico.com/news/2025/10/28/judge-los-angeles-t...
They've repeatedly violated court orders to either return immigrant detainees or release them. "This is one of dozens of court orders with which respondents have failed to comply in recent weeks.": https://www.cnn.com/2026/01/27/politics/patrick-schiltz-judg...
His targeting and shakedowns of Universities, law firms, and media companies is transparently illegal jawboning.
Everything about the tariffs is obviously illegal which he confirms every time he opens his mouth since he's relying on 'national security' justifications to issue them without Congress and he keeps insisting they're punishment for some random perceived slight.
Some sillier things like renaming the Kennedy Center -- the law that established it literally said that it couldn't be renamed without Congress -- so Trump firing everyone on the board and then appointing a bunch of his flunkees to vote for the name change doesn't cut it.. https://beatty.house.gov/sites/evo-subsites/beatty.house.gov...
It's a literal onslaught of illegality so I can't tell if you haven't read a news article since 2025 or if you're trolling.
(not op). I use a dumb feature phones that can receive SMS for far less than 40€ (even cheaper, but I like the music player and some other things like bluetooh for headphones). I have a "twin SIM", i.e. my mobile carier gives me a SIM with the same phone number and if somebody calls me, both phones (smartphone and feature phone) ring. SMS can only be received at one number, but you can switch the SIM for SMS reception using the carrier website. Since I only take the feature phone when I leave home (to enjoy outside time without distractions) I usually don't turn it on.
Scientists use pairs of satellites to map the small variations in Earth's gravitational field. It is possible to see groundwater depletion and changes in distribution of glacial ice, among many things.
The primary challenge in determining the mass of Earth is actually measuring the gravitational constant, G, itself. Everything else involved is known at much higher precision. The product of G and Earth's mass is known to two parts in a billion, but the uncertainty in G is ~22 parts per million.
LISA is primarily sensitive to time-varying gravitational gradients on timescales of a fraction of a minute to a few hours and won't be terribly useful for determining the orbits of objects in our solar system. (but it is very, very cool).
It is likely that many airlines accommodate people who are being deported as regular passengers on outbound international flights.
For movement of humans at industrial scale, though, there are only a few operators with ICE contracts. Avelo, GlobalX, Eastern, Key Lime Air, and Omni come immediately to mind. For international flights, there's at least one Learjet operator that flies a bunch for them, too.
True. Applications as these go back a few decades. From the news buzz when it was launched, ASCI White was similarly used in early '00s to understand nuclear explosions and shockwave propagation (instead of relying on live tests) - classic CFD problems[1] Successor supercomputer clusters were also used to do weapon design & nuclear physics simulations. One supercomputer IIRC even simulated a tornado genesis.[2]
I can only imagine the classified applications must have grown ten-fold in complexity in the interim.
Yeah I was thinking they probably picked a public use case that looks a lot like their classified workloads...
The article mentions the simulation ran quickly, the time spent was debugging. Suggests to me the real classified system will be much more capable.
When one person is able to do too much too quickly, they can create more liability than they can accommodate if something fails.
It is essential that a human is responsible for the utilization of any AI output in the real world, but that is not enough. For our own sakes, we must find ways to minimize the tech-debt bankruptcy blast-radius of those who would utilize (knowingly or unknowingly) AI to create flawed systems upon which others rely.
An example: Jim vibe-codes an extremely popular micropayments app. He hires a few people and sees the company as the WhatsApp of money -- a few engineers and some agentic support staff. It pulls in a few million in VC money -- enough to draw in tens of millions of users. One day, a flaw in the infrastructure causes all of the users' unsalted banking information to be released.
Agentic AI allows that entire list of customers to be exploited rapidly, so the losses for society are in the tens of billions. Jim's company is immediately bankrupt, of course, but there are only a few million dollars to go around.
Today, most of Jim's incentives are to go ahead and build that app. The same is true for his few employees and a small VC contribution. There's not much capital at risk compared with the societal exposure.
How do we ensure that AI users are accountable not just for their actions, but for the size of the risk-exposure that they create?
reply