It does, but it's very poorly documented and quite unstable (on purpose i think). What the other commenter said about the VSCode BYOK seems to be the more reliable way.
I tried adding a Foundry LLM as Github Copilot custom model and failed miserably. But with VSCode BYOK (and Github Copilot as the interfact) i did get it working, and i can now use Deepseek V4 Flash with Copilot.
There are some product recommendation websites, Twitter, or Reddit forums. Plus, it takes time for people to find out about it - you need to keep showing up on various platforms.
Continuously and deeply reading and understanding high-quality open-source code repositories - these resources are easy to find, but they require patience. With the help of LLMs, this process will be much easier than before.
Parmesan chicken, gulasch and spatzle, vegetable soup, summer tortellini skillet, lasagna (~3 hr), baked mac and cheese, baked spaghetti. I usually cook some frozen vegetables if the dish doesn't have any.
I've had similar thoughts before, and I've seen some frameworks used to assign different tasks to various agents so they can work together to complete a job. But let me be honest, with the improvement in model capabilities and the increasing cost of maintaining these frameworks, there's less and less room left for such tools to survive. The best solution might be to use a single agent and train its preferences over time.
I do agree with you, and the agent ecosystem is quite different today, compared to when I first started building this with Opus 4.6 and GPT-5.3.
But still I don't want to bet on one runtime as a single point of failure, and different vendors have different strengths (and outages / rate limits). That’s why I am building this in an agent agnostic way.
I presume OP is referring to Microsoft Windows Vista, released at the end of the year 2006.
It had a poor reputation. My vague recollection is that I experienced an irritating number of driver-related BSODs. To me, it now feels like a standard Microsoft experience of them pushing out a new OS then fixing it with service packs over a few years.
To answer OPs hypothetical thought-experiment:
In my view, running Windows Vista today would firstly have the problem that Microsoft has added lots of extra stuff (OS APIs, runtime libraries etc) and you would find that quite a lot of software won't work without stuff in newer versions of Windows.
I feel that if someone added enough extra stuff to Vista to run current software to Vista that it could run most current software then it would not really be Vista anymore.
Most of the Window XP source code has leaked. As far as I know, no pirates are patching, recompiling and releasing versions of it.
Patching Windows security holes is a lot of work. I think Vista was vulnerable when some NSA tools leaked in the year 2017 and it was possible to remotely seize control of Windows machines if port 445 was reachable and things like the wannacry worm caused global chaos.
I'm fairly sure that when Vista reached EOL it still had a lot of LPE vulnerabilities.
It's model-agnostic, so anything works. As long as the model is good at instruction following, it won't have issues with using the tool (and the AGENTS.md workflow).
Our company also requires everyone to use more AI-related tools, and I don't think there's anything wrong with that. But the quality of work produced using these tools really depends on the individual's ability. Some people don't put in much effort, and the results they produce are really sloppy, which bothers me a lot.
I think there is actually something wrong with that. What should matter is the work produced, not the tools used to produce it. If AI tools really are all they are cracked up to be, then people using them will get ahead, and the company can justifiably point out "your peer gets twice the work done as you" to the other employees. But mandating tool use in and of itself is senseless and counterproductive.
Arguably, what should matters is the impact your work has on your end user.
I agree with you that it's up to the people to adopt new tooling so they don't get left behind.
But from the company perspective, it's beneficial to get people using the new tools as fast as possible so that the overall impact on end user is accelerated.
An example is the adoption of CNC machines in factories. If you're a old timer using mill and lathe, you may argue that the company should not enforce all its workers to learn how to use CNC machines because you can still use the lathe and mill just fine. But from the company perspective, having you learn and run 3 CNC machines at the same time is much better than have you run 1 mill or lathe.
The unfortunate reality is that unless you use these tools it is impossible to keep up. People using these tools well are substantially more productive, and often they were already the most productive.
