Goodbye, $165,000 Tech Jobs. Student Coders Seek Work at Chipotle.

[megatronprime]

Once upon a time manual machinist said that exact thing about CNC machines. Carburetor mechanics about fuel injection, etc.

Yes AI code is still not at the level of top programmers, but at the rate it has progressed in just a couple years, in a few more it will be better than 99% of top programmers.

We still have manual machinists in 2026 though, a of people still learn machining and work on a lathe or mill for a living, even though there are cnc machines, cnc machines have a high cost too

[Newbeeee™]

ChatGPT is certainly iffy....
Certain things great, but my experience is the exact same Q in the exact same browser but on 2 different pc's, can give totally different answers.
I asked a base Q about a 6 bolt cylinder head tightening sequence and it said centre out then produced a diagram of this

5  3  6
1  2  4

I pulled it up, it agreed it was duff info, and then repeated.
Same Q on another pc and straightaway the correct answer.

Also, in searching for info scraping specific forums (reddit HD etc) for specifics, such as converting a harley front end from 1" handlebars to 7/8" it comes out with shit such as machine the 1" clamps out to fit the 7/8" bars....
 

[Here's Johnny!]

ChatGPT is certainly iffy....

Stop asking it stupid questions?!?!?!?!         

[YoDoug]

The 200$/month chatgpt plan for business is good, you get fast inference and u can make long ai videos and such, capable ai agents, Sam Altman says open ai loses a lot of money for each account, because people are using it much more than they anticipated.

They have an api too, so u can embed chatgpt into any app and just pay for tokens that it needs to generate.

For coding there is a new app that is popular right now called cursor, I would look at that one for sure based on reviews but i havent tried it.

They recently faced backlash for a large hike in prices

We are getting ready to drop our OpenAI subscriptions and move everyone to Gemini. Gemini outperforms Chat in almost everything. Plus they have soo many extras, like NotebookLM.

I didn't like Cursor. I use Google Antigravity with a Claude code agent running in it. That seems to be the best for my work flow. FWIW, Cursor, Antigravity, Replit and a few other AI enabled IDE's are all built on top of VS Code so they all run very similar. 

[YoDoug]

We still have manual machinists in 2026 though, a of people still learn machining and work on a lathe or mill for a living, even though there are cnc machines, cnc machines have a high cost too

Go try finding a job as a manual machinist. There may be jobs that require the knowledge of manual machining as part of the work, no one is hiring manual machinists.

I never said AI will replace all programmers. I said there will still be a need for high level tech people and AI managers.

[megatronprime]

Go try finding a job as a manual machinist. There may be jobs that require the knowledge of manual machining as part of the work, no one is hiring manual machinists.

I never said AI will replace all programmers. I said there will still be a need for high level tech people and AI managers.

There are more than 10 jobs i could get as a manual machinist withing 15 minutes from where i live if i wanted to.
It's very easy to get a job if you know conventional lathe, very easy

[megatronprime]

I have experience too, I make a mean knurl!

[Here's Johnny!]

Go try finding a job as a manual machinist

We have been trying to recruit one for over 5 years. We will be hiring an apprentice and training one it seems.

[YoDoug]

There are more than 10 jobs i could get as a manual machinist withing 15 minutes from where i live if i wanted to.
It's very easy to get a job if you know conventional lathe, very easy

Dude, where do you live, 1987?

[Bucky Cornstarch]

Dude, where do you live, 1987?

I swear when I lived in Michigan over these past 3 years it was 1990. They're gonna flip when they hear grunge!

[TylerBeer]

I'm not very optimistic about the distant future. In the entirety of human existence people have always been greedy. It's not going to change because of AI. In the future we are going to see a massive shift. Currently humans produce value in 2 ways, production and consumption. That will change to the majority of humans being mostly just consumers. That being said, the only way billionaires get to become trillionaires is to take away all things that don't contribute as consumers. Homeownership, automobile ownership, 401k's, will be only for the few in society that still are needed for production, IE, the AI managers and tech people. The high UBI that Musk keeps spouting off about will only be high enough to make sure you can consume enough to keep the elite's wealth growing. You can rent your housing, transpo, etc. from them. It will also come with a lot of strings like digital currency, completely tracked life, etc.

The near future for work will be a slow change. There are roughly 160 million people employed in the US. AI and robotics will take a good portion of those jobs over the next decade, but it will take some time to take the majority. Small businesses will be slow to adopt. The ROI and logistics of implementing humanoid robots and other AI driven advancements will take time to reach levels acceptable for small business. UBI will not really take hold until we start seeing unemployment in the 20%+, until then the economy will sustain although the middle class will shrinking will accelerate. When UBI does start the remaining jobs will still pay and offer a slightly better life, but because of the amount of competition for those jobs, pay/benefits will be capped by market dynamics. A lot of people say learn a trade instead of college or white collar or tech work, the problem is, the US doesn't need 20 million plumbers.

More specifically to your question, in programming type tech jobs, there will be a transition for some time. Currently there is way too many workers. You will see wages and benefits slowly degrade as the competition for remaining jobs increase. Far less people will start seeking education in those fields. Eventually it will balance out as older experienced workers leave the workforce those jobs will open up to the remaining workers. It is going to be rough for a decade or two.

I have told my kids this scenario for the future. They ask what can they do. My advice is while they are in their 20's, do everything they can to get $100K invested. They can't wait until their 30's or 40's to think about building wealth. If it means they live at home until they are 30 and/or work multiple jobs and/or live like they are poor, it is imperative they start to build that base now while the still can. To get a little metaphoric about, JFK said a rising tide raises all boats. The issue is we need to start looking what exactly is a boat and what is the tide. Employment and income can no longer be considered the boat and the tide. Your investments are your boat and the market/economy is the tide.

For me personally, that is why I am not as active on here. I really have stopped following politics and tried to remove more distractions from my life. I spend a lot more of my free time now learning and studying. I am working on building my consulting business. I have enough in my 401k/investments that we will be able to retire and not be dependent on gov, but I also know it could be better. I don't know if I will figure out a way to build or at least start real generational wealth over the next decade or so, but I know it won't happen if I waste my time following politics and arguing with people I never met about shit I have no effect on.

Lastly, as a guy who works with robots, I see a massive opportunity in Raas (robots as a service) coming soon. In the early days of the humanoid robot takeover, most small businesses will not be ready or tech qualified to integrate them. I think there will be an opportunity for third party integrators to have fleets or robots for lease with integration teams to get them set up. this will be some of shifting work type for out of work programmers and tech. Some one will need to come in asses job functions, create integrations, API work, and bridge applications to get these robots tied into a company's systems and processes. Develop training systems and processes for the robots. Validation, and testing, etc.

That's about exactly how I see it too except there historically new tech has always created new types of jobs that are hard to see coming - ie the smart phone industry produced a wild amount of careers that in the .com crash would have seemed like a pipe dream.

As for RaaS I went down the road a bit with a company called Greymatter who does robotic sanding/polishing robots as a service and they had some impressive stuf, VC money, but their value proposition still didn't make sense (for us anyway).
Once price comes down a bit in that area it's going to be a great option for some companies

[CNCAppsJames]

Broom Pusher checking in... 1991 or 1992. I can't remember. Didn't last long before I started turning handles. I'm nothing special.

My youngest started at G-Code's shop cleaning machines and I think coolant tanks. That lasted a couple months then he moved into the tool crib. He's smarter than I am.

My second son, he started at Trinity Automation doing basic labor stuff and installs, moved into CNC service at Selway and now he's a Service Coordinator for two regions in our territory. He's also smarter than I am.

If it can happen for us, it can happen for others.

"Just learn to code." That aged well.

Always be moving forward. Stagnation is certain career death.

TTFN

[CNCAppsJames]

Once upon a time manual machinist said that exact thing about CNC machines. Carburetor mechanics about fuel injection, etc.

Yes AI code is still not at the level of top programmers, but at the rate it has progressed in just a couple years, in a few more it will be better than 90% of top programmers.

Free fiss and that next 10% is gonna cost $1B if not more. I don't think the money is there for that. 

I've been playing with AI toolpath generation in Fusion for ~1 year. It's made exponential improvements. I judge parts on a scale of 0 to 10. 0 being sheetmetal with holes 10 being crazy inconel hydraulic manifolds with hundreds or thousands of measured features, heavy on full 5-Axis, and crucial operation order. It can handle the majority of level 5 stuff. It is still pretty low-ish on process efficiency but will make a good part. It will not satisfy people that are cycle-time sensitive. Not even remotely. Onesey-twosey customers, that's not usually gonna a show stopper for the most-part. Level 7 parts w/ 5-Axis toolpaths are where the heavy drop-off begins on AI developed programs. Levels 8, 9, and 10... those parts tend to be HEAVY on things that really have a very high number/degree of process variables that require experience that is extraordinarily difficult to quantify. The real development money will go there... or try anyway. 

#WeHaveSeenThisMovieBefore

:cough:

#STEPNC

:cough:

 

[gcode]

My youngest started at G-Code's shop cleaning machines and I think coolant tanks.

Yup, skinny ass gringo kid had it rough for a while. but he took it in stride and was my go to guy in the tool crib in short order. We were sorry to see him go, I was glad he was getting out of the PRK though. There is no future for a young man here.

[YoDoug]

Free fiss and that next 10% is gonna cost $1B if not more. I don't think the money is there for that.

I've been playing with AI toolpath generation in Fusion for ~1 year. It's made exponential improvements. I judge parts on a scale of 0 to 10. 0 being sheetmetal with holes 10 being crazy inconel hydraulic manifolds with hundreds or thousands of measured features, heavy on full 5-Axis, and crucial operation order. It can handle the majority of level 5 stuff. It is still pretty low-ish on process efficiency but will make a good part. It will not satisfy people that are cycle-time sensitive. Not even remotely. Onesey-twosey customers, that's not usually gonna a show stopper for the most-part. Level 7 parts w/ 5-Axis toolpaths are where the heavy drop-off begins on AI developed programs. Levels 8, 9, and 10... those parts tend to be HEAVY on things that really have a very high number/degree of process variables that require experience that is extraordinarily difficult to quantify. The real development money will go there... or try anyway.

#WeHaveSeenThisMovieBefore

:cough:

#STEPNC

:cough:

I guess I should first clarify, when I said 90% of top programmers, I was referring to Object oriented language computer programmers, not CAD/CAM. Those are two very different things, especially for AI. Those two things are not an apples to apples comparison for AI abilities.