But this is exactly where AI can actually be useful — so long as you don't depend on it to identify all of the design issues. It can review mountains of information and find things in it you might not have caught. You just can't assume that it will catch everything it "should" — that is, what it should be capable of catching is not necessarily what you are capable of catching.

Meanwhile, that periodic report has requirements, so you're going to have to ask it to write the report, then to review it to see if it met the requirements, then to review its own statements about whether it successfully met each requirement, because it won't do any of that sanity checking without being told to do so.

Widely? No. Divergent? Yes. Run the same query multiple times and you'll get different answers. The more input you give, the more subtle the differences are likely to be, but there's always a chance it will go completely off its nut and follow some line of "reasoning" that goes someplace utterly unrelated to what you asked for. This is less likely in special-purpose models, but never impossible.

It can't, though, and for two reasons. One, a person is required to check off on the data. Two, LLMs don't faithfully represent data. They ingest it as tokens and then what they crap out is influenced by random numbers produced internally as part of the token selection process. That means a person cannot certify that the data was faithfully reproduced without checking literally every value, because that's simply not how LLMs work.

You can potentially address this with a hybrid approach, where the LLM uses a tool to collect and manipulate the data, but this is shifting the problem to whether the LLM used the tool correctly. It might be more obvious when it doesn't, though, so it might still be a win.

What LLMs seem best at when it comes to text processing, and they are still not infallible, is summarization. But you can't trust its negatives, so it's only useful for identifying other possible positives. That's not worthless, it can help you catch things, but you can't trust it to do the parts you'd most want it to do.

I wouldn't want them to be relying on AI to do reviews though. Maybe two AI, looking only for issues and reporting them back, like you can have AI look for security issues with computer code. And a proper review done by a human.

Flock doesn't solve violent crime, it might provide evidence for a policeman to solve violent crime.

It's not evil to solve violent crime. You are getting over excited dear.

However Flock could be used for evil. And there is a consumer backlash because of that, it's the wrong product.

More interesting is why is there is so much violent crime in your culture? If you are an American, you appear to have a violent culture, perhaps the most violent in history,

the only one to have dropped a nuke.

This is when you, the accused, either demand a lawyer be present to represent you, or clam up and invoke your right to remain silent (and let them prove their case on their own).

The mistaken color is another common police tactic. t.

I saw part of the testimony. This all comes down to human over-aggressiveness—not even mistakes.

The detective used Flock to look for all Dodge Durangos within a radius of the accident. They located her at her apartment. So far, so good. But her car is a different color than what eyewitnesses said. Other details did not match up with eyewitness accounts. There was NO damage to her vehicle, but they arrested her anyway. The detective doubled down and the police continued there prosecution of her.

I'm no fan of Flock, but in this case it's simply bad police.

The basic solution is FOSS. An individual, a school or a business of any size can self-host the AGPL-licenced Elements server for free, and use their FOSS client from F-Droid. What they charge for is buzzword-features like "Corporate oversight", "Dynamic scaling". The "Sovereign" offer for governments is an airgapped deployment (internal communication of intelligence agencies)

Basically they're a service company around a FOSS solution.

Information taken from: https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Felement.io%2Fen%2Fpricing

1) The Pro version isn't open source, but purchasing their proprietary offer also supports the AGPL version of their server 2) The EU wasn't going to develop anything internally, it's not what they do; they open a tender for the supply of services. They wrote the tender to somehow exclude Microsoft, maybe they wrote "must be based on an open protocol" (Elements is based on Matrix). Any company, whether local FOSS developers or large IT consulting companies could have proposed to do what you suggest, take a fully FOSS offer and take it to the needs of a government organisation. Maybe Elements won because they have an existing offer and not only promises.

Personal piggy bank? Isn't Trump the only president whose net worth actually decreased during the presidency? The decrease in wealth is unlike most modern presidents.

Element Pro is not Open Source. So the EU is swapping a shitty proprietary solution for another shitty proprietary solution. And what will they do when the company that developed Element Pro is bought by an American company ? Swap to another one ?! Stupidity and corruption is deeply ingrained in the European Commission.

Personal piggy bank? Isn't Trump the only president whose net worth actually decreased during the presidency? The decrease in wealth is unlike most modern presidents.

This could work in conjunction with a future "just walk out" checkout implementation. You see a price at the shelf, pick up the item, you're presented with a bill in your email. You can't claim the price changed on the way to the register, and they know what price they showed you.