bitwize

The high-level Apple guys were tripping balls, especially—in all seriousness—Jobs.

It's an issue because Linux is a trademark. Microsoft cannot call something a "Linux something" without infringing a trademark, but they can say something is "for Linux" the same way third-party companies could market e.g., WordPerfect for Windows. Hence Windows Subsystem for Linux, mirroring the old Windows Services for Unix:

https://en.wikipedia.org/wiki/Windows_Services_for_UNIX

It's a subsystem instead of "services" because WSL1 really was an NT subsystem: a "kernel personality" that knows how to interpret Linux system calls just like the DOS and OS/2 subsystems of early NT.

Too much rat wrestling. You could bang out a few lines of Tcl in your editor and have a working GUI. Even stub out the commands that widgets run, tweak and reload until it behaves right, then add those back for a working program. Absolutely nice for providing a visual interface for your Unix scripts or programs, or for building whole applications in if you're daring.

I think of it as an alternate take on Lisp's metaprogramming. A form of cosmic-horror take that will drive you insane, yes, but Lisp-like metaprogramming nevertheless. Stare too long into Tcl's deadlights and you will want to be there.

That's new-school VB. Tcl was doing automatic widget layout back when VB was using absolute "dialog units" for everything.

I have to concede that this should stand as a testament to the power of generative AI and what it allows ordinary developers to accomplish. It's hard to complain about "slop" when this guy magicked an 8086 graphical operating system out of the aether and then ported it to a number of different architectures, running efficiently in space and time on each.

If you still believe that writing software "by hand" isn't cooked, you haven't been paying attention.

I find it remarkable how often quotes from Orson Welles's "Frozen Peas" apply to my career as a software developer:

https://en.wikipedia.org/wiki/Frozen_Peas

https://www.youtube.com/watch?v=V14PfDDwxlE

Back in the Agile era the most applicable one was "I gave twenty times more for you than for any other commercial I've done. You are such PESTS! Now what is it you want? In your... depths of your ignorance, what is it you want?"

But in the AI era, the most applicable quote becomes: "It's full of, of, of things that are only correct because they're grammatical, but it's tough on the ear, you see..."

Cloudslop doesn't provide high availability the way high-availability platforms provide high availability. IBM is still doing brisk mainframe business. You won't find mainframes in many data centers because you need nearly a data center's worth of hardware to provide the volume and availability of one mainframe, and that can fit in one cabinet. (Mainframes also benefited from VLSI; new ones aren't dinosaurs that require three-phase power, their own floor of a building, and their own aircon infra.)

You know, I hear it's amazing when the famous stuffed purple worm in flapjaw space does a raw blink on Hara-Kiri Rock. I need scissors! 61!

That's a lotta Pepsi points esp. in this economy

https://en.wikipedia.org/wiki/Leonard_v._Pepsico%2C_Inc.

More like Mission: Impossible. Replete with a mini-doco about how Tom Cruise literally for reals skydove out of the Blackbird without a chute and landed on the enemy rocket in one take.

One of the major differences between like, your PC, and a high-availability server is that high-availability servers require those things as well—and get them, piecewise, while still running. It's like doing a complete overhaul while the plane is flying.

IBM mainframes, famously, phone home if they detect a faulty component. Service personnel will be on site, same day, to replace it before you even knew it was there, let alone had the opportunity to ask.

They might be forced to... (gasp!)... sell wine before its time.

This. AI can make sound architectural decisions on small to medium-size problems, and even root-cause and generate a well-factored fix. In one shot.

It absolutely won't be long until product managers are the only humans who actually need to be involves with the software development process.

"Sober" in this context means "having overcome a dangerous alcohol habit/addiction". Pretty standard alcoholic terminology. Furthermore, much as they might enjoy a drink, alcoholics who have "gone sober" MUST abstain, as alcoholism is seen as lifelong once acquired and the risk of lapsing into self-destructive behavior is too great after even just one drink.

"Teetotaler" usually refers to a person who abstains from alcohol on general principles (does not like the effects, taste, thinks it's a moral vice, etc.).

Tell me you're not using a frontier model without telling me you're not using a frontier model

That's okay. Reviewing the code will become the agents' job as well.

A couple more step functions in model capability of the type we've seen in the past year, and there will pretty much be no reason for humans to be involved in the development process at all. All humans would need to do is communicate clearly what needs to be made and flag problems as they come up.

> Ever have someone try to haggle a used K5 or Pentium chip that was priced at 5$ in 2002? I have.

"Five dolla? Good LAWD that's a lotta money! How 'bout I give you a nickel and you let me use your adding machine for five minutes?" https://www.youtube.com/watch?v=_yoAVDuE7XY

It was MBTA. I'm from the northeast.

"This is ook-waeid"

I love the FM synth in the music. Some of the patches from Sonic games can be heard.

When Dario talks to himself sotto voce and then answers himself in a raspy voice, Gollum-like, made me lose it. I once encountered a severely cognitively deficient individual on the bus who calmed himself in the exact same fashion. It's almost a perfectly succinct psychological snapshot of the character they're trying to create.

Wellp, so much for the rest of the jobs.

There's a mechanical tank simulator at the Swiss Military Museum that was built in the 1970s and used for actual service before being retired and becoming a museum piece:

https://www.youtube.com/watch?v=AcQifPHcMLE

The simulator has a simulated cockpit for the user, which controls an armature with a camera attached that is piloted around a miniature battlefield environment and provides a video feed to the cockpit. A "foot" on the armature senses terrain elevation changes, which are then supplied as movement feedback to the simulated tank cockpit. All of this was controlled by a 1970s mainframe, for which replacement parts were unavailable so it was replaced, in the museum exhibit, with a Raspberry Pi.

It looks like hardware emulation was NOT used; rather, they translated the original program from paper printouts to a modern programming language: https://www.raspberrypi.com/news/simulate-driving-a-1970s-ta...

The CEO vibe-coded an app last week and shat it out onto Railway, contravening the company's security, testing, and CI/CD practices. As far as he is concerned, that is proof that LLMs are a massive productivity booster, and as such must be used by every employee in the company. An agentic workflow which is to be accepted as standard and used for all future development companywide is forthcoming from upper management. Keep an eye on your calendar for the meeting explaining it.

P.S.: An employee discovered, five minutes after deployment, that click events are not properly handled in one of the app's buttons, leading to visual and behavioral glitches.

Indeed, this whole story has "farmer leaves barn door open and has shocked-pikachu-face when his horses escape" energy.

School homework is one of those things which, like the OnlyFans economy, I hope AI utterly destroys, despite my general detestation for the corrosive effects of AI on society.

When this news dropped I thought of two things:

1) the Morris worm, which scattershot a bunch of known exploits until it hit paydirt, and then used whatever it found to compromise and replicate itself on the host system;

2) a story here on Hackernews about how someone got the fuzz tester American Fuzzy Lop to "learn" how to produce well-formed JPEGs and PDFs by pointing it at a JPEG or PDF decoder; the tester can record which code paths are followed and with enough random input can find a path into the depths of the system under test... but doing so for a decoder means actually constructing what it is meant to decode.

Neither of these are particularly "smart". But a brute-forcing machine gonna brute force, and it has the potential to cause a lot of damage. If you built a Morris worm with a fuzz tester on its nosecone, think of the mayhem you could cause! If you could examine the logs you'd probably find some undiscovered vulnerabilites in there, too! Maybe LLMs can just do so more efficiently, or maybe they let people who are too ignorant to have that kind of power vibecode their own fuzz-tester-tipped Morris worm.

Hey, even Joey can hack a Gibson with enough determination.

We experienced another step function in model quality over the summer. Models are now one-shotting medium sized tasks that require significant engineering judgement. We're moving into an era where the production of software will be done entirely by machine, requiring only an abstract description of the need the system is to fulfill in business language. This has been the goal since the 1970s. "Plan mode" isn't necessary when the model can suss out what you need from a few sentences of prompt and make competent engineering decisions to fulfill that need.

Brooks was wrong. We have a silver bullet now.