... by scientists, not by Mo McBubba the HS-grad regurgitating alt-facts from some yt channel.

If this were true, we wouldn't have millions of dissatisfied young and middle-aged men feeling left behind and without dignified work.

Also, we don't need ditch diggers, that's what machinery is for. And chefs and car washers depend on customers with enough disposable income to create a sufficient demand of their services to support all those men performing those services. Which is why we need well-paid white and pink collar workers.

Well-paid white and pink collar workers provide the real trickle-down, since their discretionary spending support between 5 and 7 jobs, per capita.

We do not have too many college graduates. Anyone who says this needs to provide some sources to support that claim.

We can have an argument that we have too many for specific fields, but we are in a post-industrial service/knowledge economy that depends on knowledge workers.

With that said, I do not believe every person needs to go to college. There are trades that pay very well: welding for instance. We rarely see a welder, mechanic, HVAC specialist, electrician or plumber struggling to make ends meet, do we?

So, we need to create paths for people, regardless of gender, to get into specialized trades, which is what we need. As for the HS grad mentioned in the article, $15/hr is starving money depending on location. It might be fine for a young kid, but man, you better find a way to increase that by 50% very soon. $15/hr just doesn't cut as they get older.

Without making any judgements: If I were confident in my skills but don't like the company, but I'm confident to find a job with a comparable compensation package, I'd take the 16-week pay (4 months tend to be enough to find a job if we are aggressive about it.)

However, if I were not confident, I'd stick around and try to get through the PIP... while aggressively looking for another job and jump ship as soon as possible.

Either way, if I were to be put in that position, it would be a sign to bail out. It would just be a matter of how to carve an exit strategy.

We can say the same about teaching kids how to paint or play an instrument. The point of K12 education is not to teach specific skills, but to expose kids to a variety of topics, and to learn to socialize and work with others.

And coding is not an specialist skill. My kids learned enough coding to understand control statements and how computers work, in 4th grade I believe. I doubt they remember all they learned, but they, like their peers, are very computer-skilled, no doubt in great part by their early introduction of coding.

We are teaching kids what coding looks like. We aren't asking them to develop a compiler or a distributed e-commerce site. C'mon. The basics of coding are not a specialist skill. It's like saying basic algebra is an specialist skill.

Can you read? Can you write? Can you do enough mathematics to plan a budget, understand compound interest or calculate how much paint you need to cover a wall? Can you look at a work-related report and extract information pertinent to your tasks? Did you acquire some understanding of arts that gives you some enjoyment, which is necessary to have a balanced life in this busy world of ours?

Seems to me you are failing to appreciate what public education gave you. Unless you learned all that on your own, in which case, good for you.

But that can only be determined with a basic exposition. Talents are discovered and nurtured via exposition. Additionally, K12 isn't about creating masters of specific domains, but to produce young people that have, ideally, a sufficiently holistic knowledge of how the world they currently live in operates. That includes the basics of reading, literature, art comprehension, finance, mathematics, science and technology.

We are a technology-oriented world, so it is imperative to provide an exposition of it. Talent development only comes after that.

I think a lot of what you said is on point, yet some of it is overtly simplistic.

It is absolutely fine to dick around AI tools just as coding-oriented kids were dicking around with QBasic and code snippets from 2600 or PC Magazine 30 years ago.

Kids are supposed to dick around, or rather, experiment with tools and artifacts in their enviroment. Let them dick around with App Inventor or scripting with Illustrator, or with Canva or Python or whatever. Let them dick around with Lego Mindstorm. Let them dick around with musical instruments or electronics or debate classes or painting, or whatever.

It's absolutely great they all have the same access. I doubt kids need to be reminded of that because that's not what their focus will be. Those who like it will pursue it further and succeed. Some will be rock starts. Others will not.

And those who don't like that path will have some knowledge of how things work, and will be more proficient in using computerized tools that aren't necessarily geared towards programming.

Kids' don't need mastery in everything they are exposed. They just need exposition and a chance to experiment as they grow and find their way in this world.

Those are some of the most famous last words among software system projects. Please carry on while I grab my popcorn.

The bulk of software is written here in the USA, for starters, so it seems fair we support this effort. Additionally, by making it global, we get users from around the world to test and submit fixes for free. If we were to limit access to the US, we would lose global contributions towards security.

It's not about fairness, but strategy.

Anything these bastards don't understand is a prime target for DOGE. Remember when it laid off people critical to maintain and secure our nuclear arsenal? And what's the efficiency in laying off close to 100K people, all at once, half of them veterans (and thus you can assume they have military training and armed)?

Efficiency has nothing to do with DOGE decisions. Don't even go there.

Uh, where have you been all these years this system has been operated from the US? This has been a globally trusted system.

But that means ramping up production in other countries, and that cannot be done in a matter of months, but years.

Additionally, there's no guarantee that the tariffs will go up against other countries, or if tariffs will go down against China.

The decisions of what to sell where and what to manufacture where cannot be made in this chaotic environment. Push come to shove, companies would be ok with tariffs going, I dunno, 500%, against X or Y country as long as they knew they would stay that way and that current tariffs in other countries do not change from week to week.

They require consistency to know how to plan and budget operations for each fiscal year.

PS. I really think Apple should diversify and spread its manufacturing plants more into South East Asia, and India, and perhaps LATAM. But this cannot be done overnight, and cannot be budgeted with the current tariff mania.

BTW, Grammarly provides a grammar checker as well as a LLM writer's assistance tool.

Dyslexic here. Don't use general disabilities to excuse poor behavior. These people aren't cheating because they have a disability. They do it because they are cheaters.

You can use Grammarly to completely rephrase a piece of text in a manner that looked more polished than what we could do by ourselves if we put the effort.

That is, Grammarly is not just a spell-checker. It has AI writing assistant service, comparable (at least for writing/composition) to ChatGPT or CoPilot.

I work closely with my kids when they do their assignments, and I encourage them to use spell checking tools (those in MS Word or Grammarly) and to take note how the corrections take place (and why.)

I let them use the AI service to rephrase pieces of text, but I demand them they do the elbow grease first ("show me your work, those are interesting paragraphs, rephrase them here and here... ok, let's try Grammarly and see how you can make these sections better, and how you can integrate them back into your main/original writing.")

That is, the tool is supposed to assist you. It's not supposed to mindlessly rephrase it, which is what this kid did, and for which his idiotic parents went to bat with a lawyer. These parents suck.

I'm typing this over lunch, so sorry for any typos:

Half of my brains agrees with you about asking if we need such studies. For us, it's obvious LMs do not understand anything at all.

However, we reach our (correct) conclusion just by inferring on our own understanding of things work. We argue our conclusions, but we do not demonstrate, for obvious reasons (it's hard.)

And it is not wrong of us to simply argue rather than demonstrating it. How often do we demonstrate that 2 + 2 is indeed four, or that the square root of 2 cannot be a rational number.

We know

Sadly, our information space is saturated by half-baked opinions and takes being presented and shared as facts. Like the AWS CEO claiming most programmers will stop programming once AI reaches a certain level (no, we won't stop programming, we will simply change the nature of our professional activities.)

A rigorous study like this not only argues, but demonstrates, that what we think about the limits of LLMs is indeed correct: that LMs do not understand.

The study demonstrates and provides evidence of this. It gives evidence of the bizarre hallucinations LLMs produce as they produce their outputs. The study provides "meat" that backs our inferences of what is true or false when it comes to AI.