(Fellow ex-msftie here too; but I left for a startup almost exactly 10 years ago, and I miss how that older culture is apparently gone).
What I don't understand is where the AI irrationality is coming from: the C-suite (still in B37?) are all incredibly smart people who must surely be aware of the damage this top-down policy is having on morale, product-quality, and how the company is viewed by its own customers - and yet, they do.
I'm not going to pretend things were being run perfectly when I was at MS: there were plenty of slow-motion mistakes playing-out right in front of us all[1] - and as I look back, yes, I was definitely frustrated at these clear obvious mistakes and their resultant unimaginable waste of human effort and capital investment.
Actually, come to think about it... maybe perhaps things really haven't changed as much? Clearly something neurotoxic got into the Talking Rain cans sometime around 2010-2011 - then was temporarily abated in 2014-2015; then came back twice as hard in 2022.
-------
[1]: Windows 8 and the Start Screen; the SurfaceRT; Visual Studio 2012 with SHOUTY MENUS and monochrome toolbar icons; the laggy and sluggish Office 2013; the crazy simultaneous development of entirely separate new reimplementations of the Office apps for iOS, Android, WinRT, the web. While ignoring the clear market-demand for cloud-y version of Active Directory without on-prem DCs (instead we got Entra, then InTune).
> but it just wasn't IBM's software that people ended up buying.
Well, I mean, WebSphere was pretty big at the time; and IBM VisualAge became Eclipse.
And I know there were a bunch of LoB applications built on AS/400 (now called "System i") that had "real" web-frontends (though in practice, they were only suitable for LAN and VPN access, not public web; and were absolutely horrible on the inside, e.g. Progress OpenEdge).
...had IBM kept up the pretense of investment, and offered a real migration path to Java instead of a rewrite, then perhaps today might be slightly different?
It's been a long time since I read through my father's Asimov book collection, so pardon my question: but how are these rules considered "laws", exactly? IIRC, USRobotics marketed them as though they were unbreakable like the laws of physics, but the positronic brains were engineered to comply with them - which while better than inlining them with training or inference input - but this was far from foolproof.
There are instances of robots entirely lacking the Three Laws in Asimov's works, as well as lots of stories dealing with the loopholes that inevitably crop up.
"For now, I’ll go dogfood my shiny new vibe-coded black box of a programming language on the Advent of Code problem (and as many of the 2025 puzzles as I can), and see what rough edges I can find. I expect them to be equal parts “not implemented yet” and “unexpected interactions of new PL features with the old ones”.
If you’re willing to jump through some Python project dependency hoops, you can try to use FAWK too at your own risk, at Janiczek/fawk on GitHub."
That doesn't sound like some great success. It mostly compiles and doesn't explode. Also I wouldn't call a toy "innovation" or "revolution".
i mean it's the most bare-bones implementation without any engineering considerations
it's not something that would ever work industrially
people with code-generators they've made could do this just as fast as the AI except their generators could have engineering considerations built-in to them as well so it'd be even better
> people with code-generators they've made could do this just as fast as the AI except their generators could have engineering considerations built-in to them as well so it'd be even better
I think they're referring to the project scaffolding features that's built-in to framework tooling thesedays (e.g. `ng generate ng <schema>` or `dotnet scaffold`).
There's also the practice of using good ol' fashioned code-generation tools like T4 or Moustache/Liquid templates to generate program entity classes and data-access methods from a DB schema, for example. Furthermore, now there's pretty nifty compile-time code-generation in C# - while languages like F# support built-time type-generation.
...and these are all good tools IMO; but really aren't comparable to an LLM, imo.
There was some HN article a month back about how all types of gambling: sports gambling, prediction markets, etc. are becoming more popular. All of this presumably downstream from a felt perception of economic malaise and nihilistic sense of resignation.
Some of it is that, but much of it has a much simpler answer.
Advertising and availability. Sports gambling was limited to Vegas a decade ago, now you can lose thousands of dollars while fiddling with your phone on your toilet.
If every corner store and gas station had a heroin vending machine, you'd probably see a lot more opiate addicts, too, regardless of how well the economy is doing. And doubtlessly, their owners would screech to the skies at any attempt to curb that shit.
There are good books on this effect, like Freakonomics and especially The Psychology of Money. After reading them I understood how people could gamble even though most know it's a net negative outcome; it's psychological, not rational.
Consider the example of a person who has been trapped by a malevolent aggressor in some sort of detention they find miserable. The de facto prison that they live in is surrounded by barbed wire fences, snipers, and angry dogs.
They're considering whether or not they should try to run through those perils. A clever, urbane economist comes up to them and tries to explain to them about cost-benefit analysis, wants to talk to them about the expected value of their decision, and the chances of death and all of that, how to make a rational decision.
That person is a fucking idiot. Because, from the perspective of the person making the actual decision, it's a binary outcome: accept a life of unending misery or do something, anything, to improve it.
There is nothing at all irrational about running into almost certain death when imprisoned unjustly. And there's nothing irrational about extremely risky behavior when your current life is miserable and there aren't enough levers to pull to get out of it by other means.
For the most part, the kind of people who write those books find themselves very popular and well-supported by the kinds of people who build those prisons.
Funny you say that in your last paragraph because your example is essentially exactly the type of analogy the authors make too, so you're agreeing with them. Maybe you should read the books first before speculating on their contents and writers.
Look back in history and you'll find that easy money policies, credit bubbles, and economic chaos caused by inflation fuel all kinds of vices and resentment.
> So many people below the poverty line barely making it by and then you have this stuff and Wallstreet Bets.
I agree the juxtaposition demonstrates how callous our society is - but would you agree that these are two entirely independent things, though? (i.e. we could have eliminated poverty while still having irrationally overpriced BTC; or maybe in the future when BTC's value eventually reverts back to 2 pizzas we will still have poverty and destitution.
I don't think it's independent. I think part of the reason for the rise of gambling in all its forms of late is a certain nihilism/despair among the younger generation (one that's not really unjustified; the median under-30 citizen now has no reasonable hope of ever owning a home anywhere where they could earn enough to support themselves, much less a family, for example).
In theory that would be great, but that isn't what is happening, it's getting progressively more debauched year after year. I'm not saying Stalinism or nothing, but I don't see the people with power doing anything beyond helping out the already wealthy. It's grotesque.
Do you mean it it was behaving consistently over multiple chat sessions? Or was this just one really long chat session over time?
I ask, because (for me, at least) I find it doesn't take much to make ChatGPT contradict itself after just a couple of back-and-forth messages; and I thought each session meant starting-off with a blank slate.
Reference saved memories - Let ChatGPT save and use memories when responding.
Reference chat history - Let ChatGPT reference all previous conversations when responding.
--
It is a setting that you can turn on or off. Also check on the memories to see if anything in there isn't correct (or for that matter what is in there).
For example, with the memories, I had some in there that were from demonstrating how to use it to review a resume. In pasting in the resumes and asking for critiques (to show how the prompt worked and such), ChatGPT had an entry in there that I was a college student looking for a software development job.
What I don't understand is where the AI irrationality is coming from: the C-suite (still in B37?) are all incredibly smart people who must surely be aware of the damage this top-down policy is having on morale, product-quality, and how the company is viewed by its own customers - and yet, they do.
I'm not going to pretend things were being run perfectly when I was at MS: there were plenty of slow-motion mistakes playing-out right in front of us all[1] - and as I look back, yes, I was definitely frustrated at these clear obvious mistakes and their resultant unimaginable waste of human effort and capital investment.
Actually, come to think about it... maybe perhaps things really haven't changed as much? Clearly something neurotoxic got into the Talking Rain cans sometime around 2010-2011 - then was temporarily abated in 2014-2015; then came back twice as hard in 2022.
-------
[1]: Windows 8 and the Start Screen; the SurfaceRT; Visual Studio 2012 with SHOUTY MENUS and monochrome toolbar icons; the laggy and sluggish Office 2013; the crazy simultaneous development of entirely separate new reimplementations of the Office apps for iOS, Android, WinRT, the web. While ignoring the clear market-demand for cloud-y version of Active Directory without on-prem DCs (instead we got Entra, then InTune).