We all want that relaxed, let the "personal agent" do it for you lifestyle when it comes to things like vacation plans, buy the next book as you see i've finished the current one, coordinate my business meetings, and order flowers for the wife's birthday - all that "and the company that will bring it to you stuff" predicted back in the early 90s.

Now finaly that the tech is here that can do it...none of us (and rightly so) trust the corporations operating them, or the rest of the network, to let them actually do it.

https://developers.slashdot.or... pretty much covered my thought about job satisfaction. If all you're doing each day, hour after hour, is "code reviews", that seems to me a really shitty job.

One reason vibe-coding is catching on with smaller firms is that, well, AI is "free" to them right now. I can go download Visual Studio Code, and get CoPilot for free...or maybe I ditch that and put Claude on it. Doesn't matter. I could still use it to generate code and hey, I'm not paying a thing.

They're getting "inexperienced junior developers" for free.

But 2 problems with that. The obvious one is noted above several times: how do you get senior, experienced developers when you haven't trained any junior developers (and don't think you're REALLY training the AI on your business model and real use cases - at any moment the engine host can just toss out a bunch of institutionalized memory for cost reasons and you're fake junior developers are as dumb as when you first started.

At some point, unless they just want the job and that's that, a code *editor* is going to get sick of editing. Maybe they won't, I don't know for sure...but I sure as hell know *I* would get sick of editing junk code day after day after day. There's no joy in it. There's no satisfaction. It is just the job, nothing more.

Now the second problem: when will they stop making it free?

This crap is expensive (as anybody following news headlines goes). And small firms and individual developers aren't paying for it. The big companies are by taking out loans against their stock values (case in point, Oracle, which took a huge 1/3rd of a drop in value for doing so).

So at some point, like with every free service that gets enshitified so that it can be somehow paid for, AI code generation will be no different. They're going to have to do SOMETHING to turn this free service into something that makes money. Licenses for improvements, a limit on number of prompts per day, advertisements showing up in the comments that the code generator generates, maybe even they'll start creating code that requires you to have a commercial library license and they won't generate 'clean' (zero derivative) code unless you pay separately, and then throw in all the vendor lock-in that goes with that.

All those things that already happen for "free" AI services online like photo cleanups. They'll have to start doing it here to recoup the huge investments.

We'll see how that changes the way this stuff is seen.

At every level of speech expression, there's a corporation involved. Nobody exists on the internet without any at some point.

So maybe I leave the 'big' social media and news sites (including youtube) and just host a blog as an ISP on a dedicated domain and VM? Nope, now my hosting provider is liable. So instead I just self-host my publishing on docker containers? Nope, because then my domain name provider and/or dyn-dns could be held liable.

They'll always have some corporation to threaten at some point to take my words off the 'net, by twisting what the word 'publishing' means...and I'm not paying any of these companies enough for them to be willing to defend me.

Yes, that's a slippery slope argument. Of course it is. And we've seen it over and over that conservative overlords will follow the slippery slope. The entire set of ideas in Project 2026 is exactly that - having achieved so much of P2025 they want to slide the slope into the next steps into pure fascism.

note, that was a napkin calculation - i didn't know they started dividends as late as 2011, and I don't know the dividend per share. i just put some numbers in to give the impression, but unlike certain politicians, I'm not trying to prove a point to influence policy decisions. Thought experiment, nothing more. :)

yeah, i just did a check. Adjusting for inflation, the stock price would need to be around $154 to be comparable in value. (Assuming total volume has remained relatively constant in that time).

that is, at $80 now, that means the stock is really worth only $42 in 2000 dollars.

So it is a number that's the same, but the value of $80 if you held onto it this entire time is still far less than you had when you started. Dividends won't quite add up to that gap - if 50c per share per quarter, and you kept the cash, that's $50 per share...but that's been spread out across inflation, so you'd have to curve the value down for what the more recent 50c means vs 50c 25 years ago. I'm not in the mood to run the calculus, but i'm gonna wager it still falls a little short.

The irony of the NYTimes complaining on copyrights given their history of reporters found guilt of or complicit in plagiarism is not lost to me.

"He developed it as thought experiment because he was dissatisfied with the existing metaphor."

No. He was thinking about it because the flaws in the Newtonian mathematics, and the ways some were trying to adapt Planck's maths into the observations, just weren't matching up. The mathematics didn't fit the observations to the degree, the level of detail, that was now feasible given the technology of observational accuracy.

So he thought about what would FIT THE OBSERVATIONS. The data came first, the explaining theory after.

The metaphor was wrong, but it wasn't some personal "dissatisfaction" - they all KNEW the theory was wrong. They all knew it. They just hadn't come up with the right one. Several, many, actually tried. They're forgotten because they were wrong. We remember generally who was right...unless there's some political/historical reason to keep mentioning (Lysenko, anyone?) why they were wrong.

RFKjr's administration have been using AI to generate justifications for policies that all are hitting exactly the same problems:

* AI is inventing studies that never existed
* AI is using quotes from real studies that aren't in the studies
* AI is generating summaries of studies that are the opposite of what the study itself actually concluded

and he's referencing these AI generated summaries in congressional hearings.

I have one pain in the arse use-case for NOT wanting https all the time, and to have it use http when i ask for it.

public wifis don't always trigger the o/s (in my case, a mac) to prompt with the page that allows you to sign in (be it public password, here's my email, whatever). So when that doesn't happen, you have to to go an http page to trigger it. https urls won't do that because, well, the cert that would come up from the internal service is wrong for the domain and so the browser barfs at that. So i have to go to an http page and NOT have the browser intercept it and try to do https instead.

hell, even when a human enters the data, from the business, into 'google', that data can be wrong...as I discovered just 3 days ago when a restaurant with "recently updated hours" said they should be open until 10pm...and were closed when I arrived at 8:45.

so yeah, if i'm investing in a trip to another freakin' country, damn right i'm gonna have every scheduled event or location double-checked by contacting a human at in the know some point in the process.

As a dev in a relatively large team (divided into sub-teams), the faster we make coding tools (even if it is just internal libraries to turn 20 lines of boilerplate into 1 line of code, to turn large if-else chains into declarative look-up systems), the faster the code comes out... ...and the more QA needs to happen to verify every thing is working to whatever our level of mission-critical is.

The code is not always the bottleneck (until you get to a big new feature or a big necessary rewrite). The process of QA and regression testing is, as well as the need to spread features out to avoid overwhelming the current user-base that are just trying to get their job done and hate it when something appears that breaks muscle memory. "Damn Engineers, always love to change things." - Dr. McCoy.

Of course, this presumes your management cares about such things. Maybe for internal projects they won't and that could be an outlet for vibe and 'speed to delivery' ahead of anything else. But customer-facing code has customer-facing concerns that slow things down a lot more than just how fast your devs can churn code or churn themselves.

"USENET died when ISPs noticed few users actually used it" - it also died a bit when deja-news was bought by google, turning it into google groups. ISPs were now thinking (incorrectly, of course) that they were paying for some Google branded service...that was coincidentally getting less use directly because Google was offering a web page interface to the same data.

So it was a gradual fade-out at the ISPs initially as people started trying web-bb's (never totally caught on, and survivers like SJGames' illuminati board have really low participation for the readership, really). The fadeout accelerated when Google replaced the usage of the service AND the impression it was an open, distributed system into Google one that Google alone should be paying for.

Actually he won't, because they already are using this as messaging that the Biden administration (and the already fired BLS exec) were lying about job numbers the whole time in order to make Biden look good going into the election.

Now this is where we can't trust the revisal - is it to justify the dismissal, or is it real, or is it both?

Nothing is trustworthy anymore, to the extent it ever was in the first place.

Boosting the anonymous post that was also a reply to this. Fair Use is defined as a matter of our right to reference a section of a work for commentary purposes. In general the excerpt length is 20 seconds, but there can be exceptions for longer. While the interviews tend to go longer, his "top 10s" are all less than 20 seconds in excerpts and even they still get DMCA claims. Those are the ones he wins the most, but it is getting ridiculously costly to have to keep fighting them.