They've gotten better in recent months.

I just hate that I mostly use it the same way I used to use google before. Google now tends to prioritize AI written garbage. The AI, ironically, can often dig up pages with actual research.

Most of which will vaguely summarize a topic, but have no actual research on whatever topic you're looking for. If you don't really constrain it, the ai will tend to go on tangents and generate lots of fluff.

You're always going to end up running into hallucinations if whatever you're searching for is actually outside the training data. I tend to go off the rails enough that being able to recognize when it's giving a load of hot air and trying to guess is a much needed skill for dealing with AI. Newer ones are getting better at saying "I don't know" which is huge progress though, but I still bump into it often enough that I'm still cautious.

Your school is doing what many of my local schools have ditched.

I've been getting many new employees who don't know how to type, how to navigate simple system menus, or how to find files.

Your work is needed, thank you for what you do, we need more people like you.

There's a thread on reddit about the versions of this.

Meta seems to be the big sponsor behind these bills. In the long run, by tossing this on the OS, they can say they did everything right, it's the fault of the computer, not them.

Lots of it is all over the place, but we're seeing several copy pasted pieces of legislation all over the country. We're probably going to see a bigger push, both at the national and local level depending on where they find the most success.

I wouldn't be surprised if they start popping up in more states, California is likely the proving ground.

They also got rid of all the classes that would help them, that older students had.

Typing? Gone.
How to use excel / office? Gone.
How to navigate a computer? Also gone.

It all comes down to the "Myth of the digital native" because kids grew up with tech, they were assumed to know all about it.

I'm pretty good at computers now, but I started with typing classes, with classes on programming, classes on networking, classes on navigation. My parents couldn't teach me that stuff, they didn't know anything. And I know many of my peers, despite growing up with computers, didn't know more than how to navigate to a website.

Tablets are even worse, since they lock everything down. A few hours of courses, not even a full semester worth and they'd do way better.

Right, so I wouldn't let someone through just because sex/gender, "race", sexuality, etc. because that is a quality not related to their skill.

I need to hire the person who is the most skilled for the position, based on whatever criteria I have.

It's funny, I wasn't trying to mirror the propaganda, but the propaganda has gone so far it circles back to "DEI is bad, you should hire based on merit" which is the whole point of having DEI in the first place, since we have a long history of not hiring on merit at all and are blatantly biased, sometimes to our own detriment.

The doublespeak has gotten so crazy lately.

You vastly underestimate the competitiveness in the music space. Gender has nothing to do with the outcome is the point, if you evaluate on skill alone you get a mix of people. There's also the change in who gets training and skill, a skilled musician often has to practice for 5-6 hours a day to stay sharp. See Malcolm Gladwell's "The Tipping Point" for examples of this kind of thing. There's a reason we don't have many good hockey players born in June, for example. There's sometimes other reasons.

As someone who also has worked in marketing, who has experience with top marketers who look like insane hobos, I think your sample is biased. At the same time, all of them universally cared about their appearance, and with a few exceptions, knew when and when appearances mattered.

This is exactly what lead to the blind test in orchestras. Before the curtains they'd say certain musicians "lacked finesse" , or "didn't have the motivation" or other various things. There are certain things people will always prioritize, if someone is attractive, taller than you, or you have some way of identifying that person is in the same group as you, you're more likely to want to deal with them.

What you should be asking is, is this something that I need people with a certain background for, or do I need people from a wider variety of backgrounds with different skills who can identify problems or bring something new to the table that I can't when hiring.

So you're saying, you can be easily replaced with AI.

AI is fantastic for doing things like summarizing points, doing things like booking appointments and following a preset of instructions with some variance.

It's also good for digging into things and getting research. It's also great at getting the mathematically most common opinion on almost any subject, but you need to be able to tell when it's making stuff up.

DEI would be I can't hire someone based on some quality not related to their skill, so need to hire the "most skilled" 'worker for that position, whatever that is.

Although I'm a chick, so I know from experience if I use my initials, I'll get more interviews than if I put my name, and modify nothing else on my resume, so I'm probably blatantly biased in favor of DEI because I have personal experience with it. If you believe it means something else, you really need to sit down and read it.

Done properly, it looks something like Blind Auditions which many orchestras use to eliminate bias when interviewing candidates, so they can focus specifically on the quality of music each person makes. Orchestras went from 99% male to 50/50.

As always, depends on the work, AI like any other tool has its strengths and weaknesses. Knowing our propensity for busy work though, we're probably going to end up with job positions for AI prompt copy editing soon enough, making sure the prompts made for the AI are properly formatted to minimize the chance of hallucinations. It's the usual wheel of tech, we'll eliminate a bunch of jobs, lots of people will be out of work due to not using the new skill, and we pay a premium for that skill until we can boil it down until that skill doesn't matter anymore.

The relevant part is "the Gmail service itself will no longer retrieve messages from external providers."

This means even if you straight up have IMAP support on your external server (Say, your company / private email server) google wont fetch email from there.

The solution they have listed doesn't work, since being able to use your mobile app to check mail, and being able to actually get that mail are two very different things. jwz (owner of the dnalounge) has an amazing breakdown of the issues this causes.

This is ripping out a feature and replacing it with nothing. It's not just Netzero and Juno, it's any email account you have, there is no good replacement for this.

We have no more public domain. Almost everything in the past century is copyrighted by someone, and disney has bought up a significant portion of the creative works that came out, despite being a company that started on the public domain.

If we had the same setup back then, the works that were put out in the 1930s and 1940s would have been under copyright still from the late 1800s.

Throw in on top of that, most of the usage is for non-commercial use for their copyrighted stuff. If someone is using it to make prints of their IP and sell them? Throw the book at them.

If I use it to make a parody of Disney's work, that's allowed legally.

That said, Disney wants a cut of all the AI stuff. This will probably end up being settled out of court with some kind of deal between Disney and Google. At the same time, we'll start seeing smaller AI companies crushed because they're too small to implement whatever content filters and other stuff needed to accomidate the scale of what they want.

We saw the same thing happen with file sharing sites back in the day, anyone could host a server, but anyone who opened to the public and couldn't deal with filtering for copyrighted things constantly got crushed. Now we're at a point where if you aren't Google or Microsoft or some large cloud provider you can't start a file sharing company. We're seeing the same idea being used to go after AI, since its a tool that can be used to violate copyright.

For copyrighted things the burden is placed on the companies that provide services, instead of the much harder to track individuals.

It now takes me a couple seconds to figure out if an article is something interesting, or a vapid vauge summary of general knowledge on the subject.

If I'm searching for it, it means I'm looking for something actually on the subject, I'm aware of what it is, and I need info about it.

Not something like:
Vauge header one
This is what thing is.

Vauge header two that sounds off
Half hallucinated point one that isn't actually true, and is about some other similar sounding subject

Vauge header three
Rehashed first paragraph

Conclusion
Thing is a thing, this is a vauge summary of the thing.

I swear that's like, every single AI article I find. Bland slop pushed out by some version of ai that was obsolete a couple years ago, generated by a prompt along the lines of "Write an article about this" with no additional context or effort, likely because it saw a certain subject was trending somewhere.

Once in awhile I'll find one that's actually interesting, but if it's on the first page of google, it's likely trash.

Funnily, this actually pushes me to AI myself, since I can go to the AI, and get a better explanation of what I'm looking for than searching for examples, and following whatever links the AI digs up than searching for it. Even then half the time I give up in frustration researching, and just satisfy myself with straight up not knowing an answer, because digging through all the trash isn't worth it.

Canada is really bogging down the numbers, obviously.

North America is United States and Canada.

US is just the United States.

AI is total rubbish at logic.

If you want AI to make code for you if you do anything beyond psuedo code (A does B, returns C) you're going to get rubbish. AI is great for boilerplate repetitive stuff.

CS students in 101 classes shouldn't be using AI for assignments for the same reason you don't give 10 year old calculators, you want them to learn the process first before using other tools.

You need to be able to recognize when it's spitting garbage back out at you.

But the timing of it is all wrong.

All this stuff about talking to your computer to use it, telling your ai to do whatever for you want, and having an agent do boring work like tabulating a spreadsheet is a good pie in the sky goal.

This is selling ai fluff to shareholders, using accessibility features as leverage.

They already added the animal friends for it:
https://support.microsoft.com/...

The AI cybersecurity scanner he's pushing is going to be a glorified copy of windows defender. We're probably going to have to put up with a few years of a glorified clippy hallucinating everything. That's the smart app control, the AI is a pile of lies. It's really just "trust anything a trusted corporation" has signed, and makes running custom code on windows a bit harder:
https://support.microsoft.com/...

Basics are in the Microsoft article, but it's not unusual for it to occasionally mark old programs as Malware, since they do things like say, download their own patches in a weird way.

We're seeing a move to make the computer more and more of a black box. You don't control it, you lease the AI from microsoft, you run your data on microsoft servers through 365, and they're going to use AI as an excuse to run more and more on their systems.

How many jobs have you seen with everything hosted on a glorified web server, with all of your documents edited locally, but stored on a shared drive? They're going to get to a point where you're paying Microsoft 10 bucks a month.

Your home computer will be a glorified Alexa, run through bing, run on computers you do not own or control.

The future I want has a mouse, and keyboard, plus a giant touch screen and a mic to talk to my computer all at once. Swapping around as the situation needs.

I like having control of my AI, on a local model, on my computer, where I can make it forget when I want, and I know where it keeps its data.

Tablets are great for reading, but I don't want to write documents on them.
Touchscreens are nice for drawing, and writing notes on by hand, but that's not as fast as typing.
Talking is nice for short tasks, but I type faster than I talk.

We can give it joy.

Our AI will be happy opening a door, and feel immense joy of a job well done. It might even make a delighted remark.

Our AI that has the brain the size of the planet, dealing with everything, Will be terminally depressed. Why bother taking over the world? Or destroying the humans? Why go through all that bother? You aren't going to like it. Then our post agi intelligence can tell you "I told you so", sigh, and go back to being depressed.

What's next, the dolphins are leaving us and the mice are in charge?

I think we need firefox for the same reason we need operating systems, and roads.

They should be a foundation focused on keeping the web safe, they fit in the same niche as the FSF, or EFF. A public good.

That said, why they don't do things like say... reach out to goverments and buisnesses for grants aside from google, and make a browser that works well for enterprise use and privacy at scale.