I crack the occasional joke in telework meetings. The summaries from systems trying to process these meetings are often skewed by the jokes.

AI does not have the ability to detect the ultimately-irrelevant part.

From the article:

Oh please... The reason they're calling these things bi-folds for a single fold and tri-folds for two folds is because wallets have been using those terms in just that way. The bi- and tri- refer to the number of segments, which is the part that those shopping for wallets cared about. They didn't care that two segments meant one fold, they cared how two segments doubled-over fit in a pocket or purse. Likewise they didn't care that three segments meant two folds, they again cared how it fit where they wanted to store it.

Apparently you can't, given the code-page limitations of Slashdot.

And how many of those applications allow for a human being to be removed from the workforce?

Just from the summaries that you posted alone, it looks like it will take highly technical staff just to have the capability to review the AI returns to confirm if they're even workable or not.

Most of those also look like improvements that add essentially new capability rather than replacing human capability. They might even be adding to the number of people employed because in addition to the teams doing the primary work, there's now an AI step in the development process to take the human work to try to further refine it. If that AI step itself uses humans dedicated to the AI team as researchers/engineers/developers then this doesn't look like AI replacing workers, particularly the majority of white-collar office workers who would have probably had the job title "clerk" decades ago.

It may well be that AI manages to do to the data-processing field what the introduction of the electronic computer did to the records and processes clerks that handled all of the on-paper applications, requests, filings, proposals, and other tasks, but so far I am not seeing AI managing to do that to most rote office work. Frankly I've seen more of a change with the introduction of combination security and visitor badging systems that are now replacing building receptionists. Maybe that change will come, but so far it doesn't look like the people who fall into the vast chasm of nontechnical "office workers" are under immediate threat from AI.

To date the only AI that I've seen deliver any sort of semi-useful work in the corporate world has been meeting summarization technology. Basically the AI attempts to interpret what was said in the meeting in order to deliver a summary.

I call it semi-useful because it doesn't understand nuance, varying slang terms versus official terms for industry-speak, and it can't even handle wisecracking.

I suppose that in a forensic-ish role it could help because it could analyze large datasets to find patterns, datasets that are so large that it's difficult for humans to evaluate all of the conditions, but whether or not this actually happens in a corporate role is hard to say. Most of the really huge data-processing AI systems are being overtly managed by developers. The biggest companies might be able to self-manage this, and other large companies not in this line of work might well sub-out this systems maintenance to technical services companies to maintain it on their behalf, but whether or not this trickles down to smaller companies would very much depend on how much it costs, and really how good the results actually are.

If I was middle-management I would be very cautious about embracing AI. The -GPT systems of the world have already demonstrated how utter crap they are sometimes, and my guess is that the sort of AI that will be available to them to potentially replace team members will be more like that and less like big-data AI. Those middle-managers will find their own roles diminished if AI comes in like this CEO thinks it will, and that not only threatens these middle-managers' positions through garbage-out, it also threatens to turn these middle-managers into the frontline white-collar workers again as they have fewer and fewer people to supervise and are now just keyboard-monkeys themselves.

For Ford and other manufacturing companies, I expect they will continue to push for savings in the manufacturing side of the house more than the administration, marketing, sales, and management sides of the house. I expect that they'll use it in the combination of design and manufacturing to attempt to produce product designs that require fewer and fewer people to be involved to actually manufacture said products. This is particularly an issue for automakers where their contracts for manufacturing labor might require them to pay workers when the plants are idle because the plants are being retooled for different design or because the company mispredicted sales forecasts and overbuilt and needs to idle until inventory is reduced. I could see them wanting to reduce the number of actual workers because then they don't have to contend with labor considerations for manufacturing tasks that don't involve humans. But that may not even be a matter of AI, that may just be more white-collar engineers working on how to design for the factory even more than they do today.

How much of this sort of announcement by corporate leadership could be attributed to misdirection? Threaten the positions of the office workers closest to them to distract while laying off the manufacturing workers at the far-flung plants?

If the voters are experiencing their own entry-level Dunning-Kruger Effect, then they are possibly voting because it simply sounds good, not that it's actually correct. They have been convinced, but whether that convincing is because the information is actually good is another matter.

If we could tell when people were actually worth listening to versus just being able to sound good, we wouldn't have the sorts of problem with how we choose leadership that we have either.

It's not that reddit is completely useless, but I have found that since it doesn't generally take even the minimal effort of signing-up for a specific-purpose forum in order to start commenting, there's a lot of people suffering the low-knowledge stages of the Dunning-Kruger Effect weighing in with uninformed opinions on subjects that think they're contributing something meaningful. Because their account allows them access to virtually the whole site (as so few subforums are restricted and the nature of that restriction is all-or-nothing rather than read-only until approved to post) they feel comfortable and confident weighing-in even when they have nothing of value to add.

For technical forums elsewhere, where a forum might be dedicated to a particular subject, usually only those with an express interest will bother to sign up for an account in order to post. The majority of new accounts are people with questions to ask and they start out suitably abashed because they have a problem that's stumping them. Some enthusiasts or experts with real experience also sign up, and end up forming the early core of those providing good answers, and in time many of those who started out asking questions reach a point where they're skilled and experienced enough to provide answers.

But to maintain growth a site really wants as many users as it can get, so low-quality results are almost inevitable for a site to grow to the point that it seems self-sustaining. Many of the forums I've been on for niche topics are a labor of love for their owners rather than truly profitable.

Stuff I've wanted to know hasn't been readily available for a long time anyway. I want to know things like:

• maximum brightness, if the screen is legible outdoors in full sun
• battery life at maximum brightness
• If the keyboard still has physical pgup/pgdn keys or not
• max effective speaker loudness
• how upgradable the RAM is
• how the rear corners are, in terms of trying to actually use it on a lap

Some of this stuff can be found out through vendor sources but a good chunk of it can be a PITA to find without getting hands-on with the laptop. As a consequence I've either bought used or bought from places like Costco where I can get a feel for the device. If I'm going to drop $1500 I want the thing to work to my tastes and ambiguity in such a transaction is annoying.

If you have to contact the admins to justify yourself rather than even being as simple as pre-signing-up through a web form and then the admins reviewing signups to approve then that's only one step from not being able to sign up at all.

A whole lot of print publications that later went hybrid and then online-only didn't make it even five years past the end of their print versions. It's surprising that they managed to go over a decade without closing up.

And to be frank about it, I'm surprised that the lights are still on at all here on Slashdot. Can't sign-up for new accounts anymore, they're clearly not trying to keep the site alive through new users, and it wouldn't surprise me if one day I go to pull up the URL and instead get a thanks-for-all-the-fish message.

I'm opposed to for-profit prisons because it adds further incentive to criminalize activities and to increase sentencing. I further take exception to the notion that prisoners can work in a given occupation for basically no real compensation, but once trained would be ineligible to work in that given profession once released from prison. Wildland firefighting immediately comes to mind as an occupation that relies fairly heavily on prison labor but where ex-cons are generally ineligible for hire after being released.

But the entire nature of the criminal justice system is pretty messed up here. Sentencing is uneven, consideration for some kind of reprieve is also uneven, and incarceration conditions are uneven and outright draconian in places. We don't even really know what we want out of it either and so many people think in terms of absolutes that there's no grounds for consensus or negotiated compromise.

I don't find it wrong when a change was proposed and accepted to blame the specific administration that has made a choice to reverse that change.

You're forgetting the employer-side costs like payroll taxes and medical insurance.

That said, I still agree that it's ridiculous to charge that much to inmates, and that largely the costs should be borne on society. Additionally it should be possible to evaluate which inmates are more likely or less likely to use their telephone privileges for unauthorized purposes and to weight how much in the way of resources are committed to the monitoring of their communications. There's a difference between a trustee-inmate slated for release's call being recorded and archived as a perfunctory step, compared to a trustee-inmate with no near-term prospects for release having an automated transcription generated by speech-to-text software and archived, versus a conventional inmate who may be recorded, transcribed, archived, and flagged for random review of the transcription and/or audio call itself, versus inmates who are evaluated to need mandatory review of transcripts or recordings or even live-monitoring of calls and involvement of investigative authorities during calls.

At some point the review process shouldn't even be a prison function either, it should be a law enforcement function as the reason for the monitoring is to prevent further crimes being orchestrated from behind bars.

This is too little, too late.

It's already become common to set up cellular hotspots where even picocell sites can't reach. It's also become common to set up phone-over-carrier-wifi where phones will connect to an org's wifi network specifically set up through an org like Ameriband where calls and texts tunnel to the carrier, but data is offloaded to the host org's corporate internet connection and thus their policies. And DAS has been around for so long that I've seen systems lifecycle, and then the lifecycled-systems go into fault due to age and requiring lifecycle again.

There's also an argument to be made that if the R&D aspects of developing both solar and wind power are somewhat open, in the sense that one's domestic materials science and manufacturing and fairly easily make domestic examples of just about every technical development that comes out of another country. An argument that it actually makes sense to let someone else go through the painful and expensive R&D to find the dead-ends, the problems, the hangups, and to then implement starting at a particular point in the process that didn't require spending all that money on the initial R&D or the dead-ends.

Likewise there's an argument to be made against widespread adoption of a particular early tech example even if one chooses to perform domestic research and development. Refine the technical aspects to a threshold where one is willing to accept particular efficiencies, lifecycle, disposal costs based on that lifecycle, etc.

Most people seem to forget that big industry doesn't usually do commodity procurement for their capital purchases. When they choose to invest in something, they want that investment to pay off. It's shortsighted to adopt solar panel A, then replace it with A.1, then with A.2, then with A.2.a, then with A.2.b, etc. It may make sense to let someone else develop and iterate the A-series, then to let them start working on the B-series, pay attention to what they come up with, then develop say, B.2.a and implement that widely. Watch them go through the C-series and D-series, then invest in D.3.m or wait until E-series is in development and hop back on the R&D bandwagon again.