I've actually outsourced the first iteration of that to AI. :-)

At the end of coding sessions, I generall ask it to spin up an adversial agent whose task it is to find flaws, inconsistencies, mistakes, bugs, etc.

I'd still strongly recommend to anyone shipping AI-generated code to have a thorough human code review, but for a first pass at surface level, the AI can check itself if it has some kind of separation between the generator and the reviewer.

Personally, after some experimenting, I've concluded that:

• AI is good at implementing known algorithms.
• AI is good in creating code that flows straightforward from a design document.
• AI is good at code reviews and finding bugs.
• AI (unlike me) doesn't mind writing a few hundred tests.
• But, it tends to over-engineering simple tasks to the point of "why does Hello World need enterprise-software levels of abstraction?"
• It struggles with code that doesn't have a known solution.
• It's difficult to get it into a different lane once it started going down the wrong way.

What I use it for these days is rapid prototyping. Using AI I can make a functioning prototype of an idea I have in a few days instead of a few weeks. I can quickly test out an idea for a feature or change in my existing codebase.

What I also have been using it for a while is as a code-reviewer and for someone like me who is mostly working solo, that's an insanely valuable thing. Not just the bugs themselves, but also the ability to discuss code with someone. Basically, AI is a rubber ducky that replies.

Mickeysoft lost the plot again.

The game-changing element of LLMs and agents is that they have a non-computer-language interface - you can talk to them in plain English (or Chinese or whatever). This makes complex computer interactions available to a much wider audience than any previous computing invention. Adapting humans to the AI is getting that backwards. It's a sign you don't understand the first thing about why AI is so widely successful.

I predict this will be an abject failure and a year from now will quietly be pushed to the very back of the bottom-most drawer, never to be spoken off again (though it is getting crowded in there for M$).

Meanwhile, in the real world, everyone I've talked to about these things already has a skills file to teach the AI about their personal coding style. This is what actual humans came up with, independently several times, and it works: Adapt the AI to you. Plus we've been writing coding style guides for humans for a long time now, we've got plenty of experience doing that at least somewhat right.

The solution to that is AI. Finally, one-on-one tutoring becomes feasible.

AI to me is kind of like a self driving car... it would be one thing if the car actually drove itself and didn't require a licensed driver to sit behind the wheel and dedicate their full undivided attention to making sure the car doesn't do something stupid... yet no such thing at present exists.

You can take an Waymo in any of these cities:

• Phoenix metro, Arizona
• San Francisco Bay Area, California
• Los Angeles, California
• San Diego, California
• Denver, Colorado
• Las Vegas, Nevada
• Dallas, Texas
• Houston, Texas
• San Antonio, Texas
• Austin, Texas
• Atlanta, Georgia
• Miami, Florida
• Orlando, Florida
• Tampa, Florida
• Nashville, Tennessee

Especially in america. The most obvious thing is that the bells were there to train people to work in factories.

But it's pretty well known that the education system was built to make useful workers for the Epstein class. The reason it's sometimes doesn't feel that way is that it's modeled after the education system the Epstein class maintains for themselves which is very much not designed for that purpose. The Epstein class has a completely different education system for their kids designed to teach them critical thinking and a variety of skills associated with stealing money from the lower classes. The critical thinking is especially important to prevent their kids from being taken advantage of and having their money stolen since every member of the ruling elite are crooks and psychopaths.

This is how and why guys like Kim jong-un went to school overseas at fancy schools and it's the same reason mitt Romney went to fancy schools. They're being taught to function in a aristocracy where everyone is going to stab everyone in the back because they are all complete total psychopaths. The ones who aren't get weeded out pretty quick at that level.

Getting back on point the school system has some elements where it's designed to make smarter people with critical thinking skills but those elements are blunted because of a combination of religious parents who don't want their kids coming home with an actual understanding of the holy books and as well as the ruling class wanting pliable useful tools instead of people who can think.

Information is out there if you care to look it up but I warned you it's very very boring. If you think world history is dull try reading the history of the American education system.

...maybe this is a stupid question but ... who's going to determine if the AI gets it right, eventually?

Other AI?
I mean if AI is writing the reports and AI is evaluating the results and announcing "hey we solved it!'...who actually gives a shit at that point? Why would they honestly bother to tell us?

That seems a short route to Wall-E world where we're just sipping out SUPER GULP 1000s.

Ironic that the next story my feed is that Chief Economist at Apollo Global determines net benefit so far of all the AI investment and hype is....zero.*
https://slashdot.org/story/26/...

*in my experience, that may be charitable.

Supposedly, they weren't plantations, just pine forests where all the trees have a tendency to being the same size when viewed from a distance.
Young forests, including plantations as well as land cleared for logging or cleared by fire supports a lot of wildlife whereas the old growth forests have very few large animals in them unless there is a salmon stream or such.
The bears close to civilization are as scared of people as we are of them and rarely do more then the occasional swat. They can do quite a bit of property damage and they're surprisingly good at opening car doors.
It's only when a long way into the bush that bears are likely to consider eating people. Except polar bears, they always consider people prey.

"may be AI slop" is a nonsensical statement, because if it's slop it doesn't matter who wrote it. Whether AI did the coding is not really relevant, although it is a more objective claim.

I live just outside fast growing suburbia. Lots of forest being converted to housing etc. Also a series of bad berry crops in the mountains. There are a lot of bears right in town, deer too. The problem with the bears is the townsfolk aren't bear aware. While direct interactions between bears and people are rare with an occasional swipe at a person, a bear did grab a toddler earlier this year. Parent drove the bear away and cops shot it.
The problem is bears getting habituated to humans, used to eating garbage, dropped fruit etc. A bear can go through a garage door easily, they're also good at opening car doors. Amazing the damage a bear accidentally locked in a car can do. Anyways lots of bears end up shot by conservation officers simply by being bears. It's the shits as generally the bears are more scared of people then we are of them and with a bit of work you can keep them from getting used to eating your garbage.
But yes, bear sightings are now common, even occasionally in the big city. It's kind of neat but really you can't trust them as they're wild animals

AI though, it really took a surprising and discontinuous leap in capability with Chat GPT 3, with application across soo many (most?) recognition and prediction tasks.

Including self-driving. What besides driving consumes so many hours and lives, and potentially remediable? I think it's probably the biggest AI application of the 21st century. Looking back the transition will seem faster.

It's hard to offer protection when something bigger offers your customers protection in your stead. Who cares about Getty images in an age where you can tell Google to magic a picture into existence.

It's a very reasonable approach, so reasonable that the drives already store ECC data. Why do you think it reports "corrected" and "uncorrected" errors separately in SMART data?

Is there really a need to store multiple copies of ECC codes? At some point, why not just rely on backups. Checksumming is good, wasting diskspace because of a neutrino from space is not.

No not quite. The HDD / SSD writing checksums onto the platter only cover the write operation itself. It will only confirm what bits were written to the platter along with its immediate ECC codes are correct, and it does not confirm any failure outside the immediate write operation. An example of corruption that it would miss is a power failure mid-write. The in flight bits would be written along with the ECC code before the power goes out (thanks to capacitors, especially true for SSDs which need to then note where on the drive that data is stored), but the drive has no way of knowing if the entire block or even the file has been completely transferred.

That said I don't think this is a justification for storing ECC data. Checksums, yes, but there's no need to waste disk space on this unless you're running a very critical system. Spit out and error and let the user recover from backup.

In our math classes at university the homework was graded. Every week it was 30 questions, 10 points, for every mistake you lose a point.

Now here's the rub: The answers were in the back of the book along with a complete worked solution. The point of the homework was practice, the point of grading was to make sure the practice was done in literally any form possible. Our math professor flat out said, even if you just go to the back of the book and copy the answer verbatim you'll still be learning something.