rbanffy

It was a long migration process, but I don’t remember anyone being unprofessional in the core teams, and certainly not like this.

I heard they were researching up to 18, but most of the blades were curved and really short.

Again, the one thing I love the most about Python is the lack of drama. I don’t want to be forced to practice Drama-driven Development.

Where is the Wordpress.org board to force him out? It seems he’s harming the project and the foundation much more than his company.

I get the frustration, that it’d be nice of commercial entities that use our software to do more to support its development and maintenance, but that’s not in the license. A good project will create a healthy ecosystem with sufficient voluntary influx of resources. If it doesn’t, then maybe it’s no longer healthy or viable.

There will be a time things like Python, FreeBSD, Linux, MySQL, and so many others we know and love, will have been surpassed by new successors better adapted to that time. We’ll cherish their memory, remind everyone of their role in taking us to that future, and move on.

If the packages are generous enough, it would create an incentive not to fire them, and, that way, be actually in the best interest of the company.

If the CEO is indeed insane and incapable of fulfilling his duties, and he still controls 84% of the voting shares, all options are nuclear.

Why stopping at four dimensions?

If the events inside the box can’t influence events outside the box directly, only by changing the surface of the box, it’s fair to say they don’t happen at all, and all that happens is the change of the surface.

I feel the most interesting implication of the enclosing surface being a complete description of the enclosed volume and its contents implies the contents don’t need to exist. If you think physics as the set of rules that applied to one full state of reality can completely give it’s next state, then having redundant representations is wasteful and the interior of the enclosed volume, as the interior of the black hole, might not need to exist, or to be represented at all. If you extend this to the full spacetime we inhabit, it all would be the encoded in its boundary and, therefore, only the boundary is needed.

Microsoft has a huge market share, and immense testing and QA resources, and it’s reasonable to assume this has driven down the number of unknown issues a lot.

That said, why do we think compilers produce correct code?

I didn’t realise we were already in the sub minute range for the default configuration.

This is pretty impressive.

I miss those years.

It’s now Rust on Rails.

Rendering HTML on the server was kind of how IBM mainframes talked to their 3270 terminals, only that HTML is a lot nicer than 3270 streams and HTTP is human readable.

Server side rendering makes sense for mostly static content, but for web applications, a front end framework and a back-end API are a much better match.

> and yet he feels confident enough to steer the ship in a completely new direction

If you don’t build the product that’ll kill your cash cow, your competitors will.

> Readability and conciseness still has benefit, but the trade-off of worse performance (and less static checking) is suddenly more questionable.

There is an issue with this idea - we write software, we read it, we make sure it does what it should do, that it’s implement in a sensible and efficient way, and, most importantly, we make sure the tests do what they should do.

If you no longer write the code, it’s harder to understand it. If you don’t write the tests, you need to trust the LLM did both jobs correctly (multiple adversarial LLMs can help), but, ultimately, it’s code you can no longer claim you understand. And you can’t really be sure it does what you want until you understand it.

You can probably transpile the 8080 code of the VT-100 ROM and make a very accurate replica, if you figure out the video generator - you’ll need to translate all the MOVs to VRAM and whatever CRTC the terminal used. Or just read the memory and spit 30 frames per second to your video output.

I'm thinking of building something that emulates a VT-340 terminal. I can take MAME's emulation of the 80186, the ROM dumps, and have something that just translates the 80186 ROM code into a compilable program, with the hardware behavior informed by MAME's description (which is incomplete for the 3xx series).

Intel has done a lot of work to make their CPUs better for running inference. I'm sure people will figure out other creative misuses for the instructions. One reason to run inference in-process on a CPU is latency - setting up a GPU or an NPU to run a job introduce latency and context switches. Running it in-process avoids that, which is nice for tasks where the AI part runs for a very short time in comparison with starting the job on a GPU over PCIe.

This is one of the reasons IBM introduced an AI accelerator and added specific instructions to use it on their Telum II processor - so that limited inference can run in-process in transaction processing with latencies measured in clock cycles.

I think these instructions are for the case where inference is a minor part of the time it takes to process some data on that server - in that case the server will be more effective by having more, faster memory than having to go across a PCIe but to a GPU that will be idling most of the time, or going over the network to a dedicated system.

These instructions are part of the ACE set, so I assume they are useful for AI - there are format conversions, rounding, comparison, and, of course, operations on these formats in the vector and mask registers. Looking at the documentation, which is huge, gives me the impression this will improve quantized models the most, but this is a guess.

With all the capabilities, video and audio output, it might be an excellent platform for software emulation of older devices. There is plenty of GPIO that can be driven by the 40 MHz core while the other two work on nobler things.

> understand and suffered from these gotchas.

Compatibility was their greatest weakness and, at the same time, their greatest strength. It allowed the Cambrian explosion of the 70s and 80s personal computers, each with their own superpowers.

Same here. And I wouldn’t pass anything weird.

Currently I’m looking for a VT-330 or some other terminal with Tektronix graphics. These seem like unobtanium this side of the Atlantic.

So, if someone has a Rainbow, Z-100, AT&T Unix PC, 3278 terminal, or anything like that, I’m rescue stray computers.

How do you explicitly tell the CPU to fetch the next batch of data before you start processing the current batch so that the cache is hot when you need it? I imagine there are instructions for that.

And nobody talks about the Xeon Phi anymore…

EOL hardware is usually fine for home labs.

Shameless plug (and hopeful suggestion): every esoteric/ancient computer system deserves a page like this one (that I’ll eventually finish).

https://rbanffy.github.io/fun-with-big-computers/fun-with-vm...

BTW, I have updated the container images for VM/370 Sixpack and VM/370 CE today.

I used OpenVZ to turn a decommissioned IBM desktop (which I asked corporate IT to fill with memory and disks from the other identical machines) into a departmental development server. It was on this machine we tested the scripts that migrated half a million blogs from a proprietary tool into WordPress.

Years later, the machine was still chugging around on top of that floor's networking rack. Someone added a PostIt with "Do not turn this off".

But I have to say OpenVZ didn't do all the tricks one could do with Zones, and Zones combined with Crossbow, were able to test a huge number of networking scenarios to see how things would break.

Was I the only one who read this as Futurama?

> They could have had a larger role in x86 Unix world

They were always a hardware provider. If they could have offered credible desktops at PC prices, they could have made a serious dent in the low-end server (one step up from desktops), which is an essential stepping stone for securing midrange and above.

On the onboarding thing, Oracle Cloud didn't offer Solaris as an OS option last time I checked. And, of course, they never offered SPARC because abandoning it was one of the first moves. IIRC, IBM Cloud also doesn't offer Linux on POWER (but they do offer AIX and IBMi, as well as z/OS - and Linux - on s390x). Not that long ago I tried to convince VMSI to make a deal with cloud providers and allow OpenVMS to be as easy as any other OS on amd64.