• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

Microsoft to Replace All C/C++ Code With Rust by 2030 using AI [Debunked by source]

Oh the up side, my daily user computer is a MacBook

On the down side, Linux needs to get their shit together on gaming because PC gaming is gonna be fucked in the 2030's
I had considered looking at a really nice gaming laptop recently (really good sales on laptops with 5070 or higher inside), but I simply can't pull the trigger because of Windows 11, it's a literal cancer on everything it permeates.

I'll be sticking with my PS5 Pro, Switch 2, and M4 Max MacBook Pro (which is able to play a lot of Windows games thanks to CrossOver, along with good emulator performance).

Wake me up when Microsoft isn't shit.

….it's never, isn't it?
 
Last edited:
I admire the initiative, since Rust is supposedly memory-safe, but its probably not going to be a smooth sail in any way. Especially, with shoving the workload onto AI to prove their commitment.

The intiative to move to Rust is fine. The motivation to do so is what is suspect. This is more about promoting AI than anything else. $1 bet: we will see "written by Copilot" promotions by 2030, probably long before.

Microsoft always has an angle
 
Last edited:
The intiative to move to Rust is fine. The motivation to do so is what is suspect. This is more about promoting AI than anything else. $1 bet: we will see "written by Copilot" promotions by 2030, probably long before.

Microsoft always has an angle

Agreed*.

*This post was written by Copilot. TM Microsoft 2025, all rights reserved.
 
Agreed*.

*This post was written by Copilot. TM Microsoft 2025, all rights reserved.

I submitted my comment to Copilot:

Fewkch7tDt4RXQiO.png


jon stewart boo yah GIF
 
The intiative to move to Rust is fine. The motivation to do so is what is suspect. This is more about promoting AI than anything else. $1 bet: we will see "written by Copilot" promotions by 2030, probably long before.

Microsoft always has an angle
I would be surprised if this is an actual Microsoft initiative. Sounds more to me like the obnoxious LinkedIn chest puffing that has become so common lately. They're going to be all-in on Rust until they encounter a scenario where they need the flexibility of C/C++.

I like Rust, and I have not written an OS or a compiler since college, but I'm having a hard time seeing how this is an effective use of time or money. Reeks of dangerous hubris to me.
 
I would be surprised if this is an actual Microsoft initiative. Sounds more to me like the obnoxious LinkedIn chest puffing that has become so common lately. They're going to be all-in on Rust until they encounter a scenario where they need the flexibility of C/C++.

I like Rust, and I have not written an OS or a compiler since college, but I'm having a hard time seeing how this is an effective use of time or money. Reeks of dangerous hubris to me.

You very well could be right. Would not be surprised if we hear this whole thing dies because Nadella or Amy Hood found how much time and money it would actually cost, AI be damned.

I don't know how many times I've shot down calls to move to <insert language/tech here> just because someone wanted to use the shiny new tech. If they were going to re-architect Windows from the ground up and decided to use Rust then that would make perfect sense, but to do so for the sake of doing so, or worse, because marketing said to do so, it just idiotic. The company is run by bullshit artists, top to bottom.
 
You very well could be right. Would not be surprised if we hear this whole thing dies because Nadella or Amy Hood found how much time and money it would actually cost, AI be damned.

I don't know how many times I've shot down calls to move to <insert language/tech here> just because someone wanted to use the shiny new tech. If they were going to re-architect Windows from the ground up and decided to use Rust then that would make perfect sense, but to do so for the sake of doing so, or worse, because marketing said to do so, it just idiotic. The company is run by bullshit artists, top to bottom.
Reality Reaction GIF by Married At First Sight


^ to Microsoft.
 
I had considered looking at a really nice gaming laptop recently (really good sales on laptops with 5070 or higher inside), but I simply can't pull the trigger because of Windows 11, it's a literal cancer on everything it permeates.

I'll be sticking with my PS5 Pro, Switch 2, and M4 Max MacBook Pro (which is able to play a lot of Windows games thanks to CrossOver, along with good emulator performance).

Wake me up when Microsoft isn't shit.

….it's never, isn't it?
I am hoping in 2027 AMD is going to have RDNA 5 laptop GPUs and APUs. And supposedly Nvidia is working on better Linux support.

However with this AI shit, you never know.
 
I am hoping in 2027 AMD is going to have RDNA 5 laptop GPUs and APUs. And supposedly Nvidia is working on better Linux support.

However with this AI shit, you never know.
I don't trust AMD or Nvidia to follow through, they're in bed with the other AI companies and funneling money between each other to make the AI bubble larger.
 
You very well could be right. Would not be surprised if we hear this whole thing dies because Nadella or Amy Hood found how much time and money it would actually cost, AI be damned.

I don't know how many times I've shot down calls to move to <insert language/tech here> just because someone wanted to use the shiny new tech. If they were going to re-architect Windows from the ground up and decided to use Rust then that would make perfect sense, but to do so for the sake of doing so, or worse, because marketing said to do so, it just idiotic. The company is run by bullshit artists, top to bottom.
The issue here is that Nadella is 💯 all in on AI. He is sacrificing everything including potentially Enteprrise cash flow from frustrated customers.

Enterprise Support, stability of the cash cows Azure and 365, development of things from SQL to SharePoint are all taking a punt. So I wouldn't be surprised if he himself buys into the bullshit. And CFO buys in because of promised personnel cuts.

This could severely affect MS down the road but the execs won't care since they will make a ton of $$$ meeting their metrics on AI so they justify mega payouts to the board.
 
Last edited:
I doubt the AI bubble will last another five years. Skepticism and rejection of anything involving AI beyond ChatGPT are becoming increasingly common, and even ChatGPT is viewed negatively if someone brags about using it.

Given the current landscape, promoting that your OS or software will be mainly written by AI is a tough sell. But anyway... I guess Microsoft has to defend its strategy after that absurd investment in AI so that shareholders don't run away.

All I can say is good luck with that.
 
You very well could be right. Would not be surprised if we hear this whole thing dies because Nadella or Amy Hood found how much time and money it would actually cost, AI be damned.

I don't know how many times I've shot down calls to move to <insert language/tech here> just because someone wanted to use the shiny new tech. If they were going to re-architect Windows from the ground up and decided to use Rust then that would make perfect sense, but to do so for the sake of doing so, or worse, because marketing said to do so, it just idiotic. The company is run by bullshit artists, top to bottom.
I've been in that situation many times as well. Someone wants to go all-in on the buzzword tech because they only see the benefits, but they haven't taken the time to evaluate the risks. Plus they're too amazing to maintain someone else's code. More than cost, the risk of failure from a full re-architecture is going to be too high for the people running the business to stomach. Eventually someone higher up will temper the notion with a dose of reality.

Microsoft is so full of people like this. I can't tell you how many hot-shot Microsoft engineers and architects I've had arguments with over the years who tried to box me into the latest thing that they were pushing only to see MS completely abandon the hot tech shortly after. My philosophy has always been to keep it simple and stay off the bleeding edge. Less risky and rarely fails me.
 
The issue here is that Nadella is 💯 all in on AI. He is sacrificing everything including potentially Enteprrise cash flow from frustrated customers.

Enterprise Support, stability of the cash cows Azure and 365, development of things from SQL to SharePoint are all taking a punt. So I wouldn't be surprised if he himself buys into the bullshit. And CFO buys in because of promised personnel cuts.

This could severely affect MS down the road but the execs won't care since they will make a ton of $$$ meeting their metrics on AI so they justify mega payouts to the board.

It's ok if all their other products have problems because now you have copilot to help you figure out why.

Think About It GIF by Identity
 
I have experienced more M365 and Azure bugs this year than 2-3 years prior. it's beyond ridiculous.

Same.

Excel is dog shit and the bane of my existence at the moment. Power BI isn't much better. The latest version of Outlook is also shite. So bad that I've gone back to the older version, as have all of my team.
 
As someone who has been programming in tech for 25 years, this is a really bad idea. Does it drive new revenue? No. Could it introduce a lot more problems? Yes. So the risk/reward is off the charts to the negative. This is something a division would do if they have too much time on their hands and a boss who is gunning for a promotion. Regardless of whether this works or not, he's going to claim it was a huge success.
 
Same.

Excel is dog shit and the bane of my existence at the moment. Power BI isn't much better. The latest version of Outlook is also shite. So bad that I've gone back to the older version, as have all of my team.
The integration of copilot into everything has been breaking all their stuff. It's so hamfisted in M365. The only useful integration of copilot I've seen from Microsoft is GitHub copilot into VSCode. The agent mode is extremely useful, if not highly prone to hallucinations.
 
I've been in that situation many times as well. Someone wants to go all-in on the buzzword tech because they only see the benefits, but they haven't taken the time to evaluate the risks. Plus they're too amazing to maintain someone else's code. More than cost, the risk of failure from a full re-architecture is going to be too high for the people running the business to stomach. Eventually someone higher up will temper the notion with a dose of reality.

Microsoft is so full of people like this. I can't tell you how many hot-shot Microsoft engineers and architects I've had arguments with over the years who tried to box me into the latest thing that they were pushing only to see MS completely abandon the hot tech shortly after. My philosophy has always been to keep it simple and stay off the bleeding edge. Less risky and rarely fails me.

It seriously feels like all the Microsoft support guys I talk to are trying to upsell me on shit rather than help with whatever we are working on. I try to tell them I'm not the guy that buys shit so they are wasting their time. Incredibly annoying.
 
Are these people *literally* retarded? I know someone always likes to come along and be like "no, you don't understand! these are some of the brightest minds in tech!" But I just don't fucking see it. That might have been the case 20 years ago, but there's absolutely no chance of it now.
 
This should surely push the gaming market into GabeN's loving linuxy hands... at the very least we'll have to dust off our autoexec.bat and config.sys and try to make a go of it in DOS.

MS Windows Rusty AI Edition, lmfao good luck.
 
The integration of copilot into everything has been breaking all their stuff. It's so hamfisted in M365. The only useful integration of copilot I've seen from Microsoft is GitHub copilot into VSCode. The agent mode is extremely useful, if not highly prone to hallucinations.

I don't see Copilot in Sql Server Management Studio yet, but that software went from one of my most reliable tools to the most buggy shit on my computer. I'm sure Copilot will emerge in one of the updates to prove me right.
 
As someone who has been programming in tech for 25 years, this is a really bad idea. Does it drive new revenue? No. Could it introduce a lot more problems? Yes. So the risk/reward is off the charts to the negative. This is something a division would do if they have too much time on their hands and a boss who is gunning for a promotion. Regardless of whether this works or not, he's going to claim it was a huge success.
Rewriting something that works great in the "hot" new language for nebulous reasons is such a rookie mistake. So it makes senes that Microsoft is doing it, considering they are far and away the worst big tech company on the planet. Talk about a company living on mere inertia alone. Everything they make is total garbage.
 
I don't see Copilot in Sql Server Management Studio yet, but that software went from one of my most reliable tools to the most buggy shit on my computer. I'm sure Copilot will emerge in one of the updates to prove me right.
I've been using it in preview for a couple of months. I had to provide it with an API key and endpoint from my Azure AI subscription to get it to work. It wasn't natively integrated.
 
I've been using it in preview for a couple of months. I had to provide it with an API key and endpoint from my Azure AI subscription to get it to work. It wasn't natively integrated.

Sounds lovely. Early next year, I gotta run Sentinel logs through an Azure AI workflow to try and make sense of the data flood for the security guys. Not looking forward to it.
 
I imagine Microsoft is just a bunch of indian interns and LLMs trying to make sense of codebases written by better people who have been gone for a long time.
 
Sounds lovely. Early next year, I gotta run Sentinel logs through an Azure AI workflow to try and make sense of the data flood for the security guys. Not looking forward to it.
That sounds interesting. I've been more successful doing analysis using Python tools to orchestrate workflows against models using Azure AI endpoints than using native Azure tooling.

I needed to do analysis of legacy code to extract business rules so I created embeddings that I stored in a local vector database, created a RAG app that I could use to ask questions of the code, and then used langchain to orchestrate the prompts I needed to answer questions. I used Azure AI services for model access. It's relatively cheap to use their API. It was more cost effective than trying to buy hardware to run open source models locally.

The Azure tooling is getting better but at the time it wasn't working well enough to justify the cost. The open source community has been awesome with tools.
 
You have no idea how much I hate having to work with Teams, Onedrive and Office 365....
Friends Hug GIF by MOODMAN


I do… my best time with them is answering their customer feedback survey in Teams and letting the anger flow 😂.

Some companies do not have SharePoint / OneDrive setup and use Teams to manage files… oh the pain when it of course goes wrong 😂.
 
That sounds interesting. I've been more successful doing analysis using Python tools to orchestrate workflows against models using Azure AI endpoints than using native Azure tooling.

I needed to do analysis of legacy code to extract business rules so I created embeddings that I stored in a local vector database, created a RAG app that I could use to ask questions of the code, and then used langchain to orchestrate the prompts I needed to answer questions. I used Azure AI services for model access. It's relatively cheap to use their API. It was more cost effective than trying to buy hardware to run open source models locally.

The Azure tooling is getting better but at the time it wasn't working well enough to justify the cost. The open source community has been awesome with tools.

You are way ahead of me. I've been messing around with Azure AI at a code level. Created an AI client that took business data serialized into json as a data source and was able to answer questions about it. Pretty simple stuff to do which is what I love about C#. I'll talk shit about MS all day, but they still kick ass as far as integrating new tech into development environments. Hopefully that won't change.
 
If they succeed, it's only a good thing. They kinda have to go this route for memory safety's sake.

...The thing is, it's an unrealistic goal, for now. But it's not a binary thing, if they can get to a 20-30% rewrite by 2030, it'll be nice for security.
 
Last edited:
You are way ahead of me. I've been messing around with Azure AI at a code level. Created an AI client that took business data serialized into json as a data source and was able to answer questions about it. Pretty simple stuff to do which is what I love about C#. I'll talk shit about MS all day, but they still kick ass as far as integrating new tech into development environments. Hopefully that won't change.
I have to agree with you. I have loved MS dev tools since the VB IDE and the first version of Visual Studio. I was not a fan of rolling windows apps by hand in C++, so I used an IDE on OS/2 Warp to create UI and then switched to Windows 3.1 to compile for Windows. I was thrilled when Visual Studio rid me of the OS/2 dependency I had. Ah, the good old days...

If you haven't tried it would recommend giving github copilot in VSCode a go when you start building your workflow app. Get your company to spring for the pro+ for $40 per month to give you more premium requests and let Claude Sonnet and Claude Opus assist you. You pay like 4 cents per request over your monthly allotment of premium requests but it's worth it for the time savings, imo. The agent mode can fully control the VSCode workspace, read files in it, create code, perform testing, and write documentation for you when you're done. It can access the output in the powershell terminal to diagnose issues when debugging. You can also use it in Visual Studio, but I've come to really appreciate VSCode and how well everything just works together there.
 
I have to agree with you. I have loved MS dev tools since the VB IDE and the first version of Visual Studio. I was not a fan of rolling windows apps by hand in C++, so I used an IDE on OS/2 Warp to create UI and then switched to Windows 3.1 to compile for Windows. I was thrilled when Visual Studio rid me of the OS/2 dependency I had. Ah, the good old days...

If you haven't tried it would recommend giving github copilot in VSCode a go when you start building your workflow app. Get your company to spring for the pro+ for $40 per month to give you more premium requests and let Claude Sonnet and Claude Opus assist you. You pay like 4 cents per request over your monthly allotment of premium requests but it's worth it for the time savings, imo. The agent mode can fully control the VSCode workspace, read files in it, create code, perform testing, and write documentation for you when you're done. It can access the output in the powershell terminal to diagnose issues when debugging. You can also use it in Visual Studio, but I've come to really appreciate VSCode and how well everything just works together there.

Nice. Thanks for the tips.
 
I've been using it in preview for a couple of months. I had to provide it with an API key and endpoint from my Azure AI subscription to get it to work. It wasn't natively integrated.
They have Copilot in Azure now. I was talking to one of MS architects during our meeting and we were asking him basically "what's the point". He went on about how you could ask Copilot to create a VM with parameters you want.

After a question of likelihood of a hallucination and what is even the point if pretty much everyone does deployment through automation tools, or Terraform or Bicep, the answer was it could maybe at least answer some questions about the environment.

It's all so pointless and yet MS is spending resources and reducing other areas like actual support for Azure. I am so tired of this shit. For all my dislike of Bezos and Co, AWS is miles ahead both on functionality and support side.

Anyways, I do admit some of the Foundry crap is actually useful. We have deployed agents through it and visibility and audit logging are very useful.
 
Sounds lovely. Early next year, I gotta run Sentinel logs through an Azure AI workflow to try and make sense of the data flood for the security guys. Not looking forward to it.
That actually isn't too bad all in all. Stupid naming convention aside, the security copilot can help with making some sense of that. Of course it's still all worse than Splunk.
 
Yep, LinkedIn blubbering, as expected. If they want to create code migration tools they should probably focus on mainframe stuff like COBOL and FORTRAN before C/C++. Thats where the real legacy code issues are. There's a C compiler for practically everything, so even if you don't understand the code at least you can build, link and run it pretty much anywhere. Options are way more limited for code as prolific as COBOL.
 
Top Bottom