Push Chinese models. Deepseek, kimi, qwen etc.
No datacenters locally and local oligarchs won’t have that much power over us. They won’t be lobbying that much against humans if local oligarchs can’t make money
Yeah, empower chinese models, instead! That’s the solution!
You are not empowering them if you run them locally
Doubt.
How is running a local model empowering anyone else?
I would argue - fucking empower them. Less money for us AI companies, less human firings.
Stall and Delay them for as long as we can.
Don’t chatgpt, deepseek it.
Why would anyone use Copilot when you have so many other options. Cursor just released Composer 2.5 and it’s actually decent. I should have a job right now doing this.
Because subsidized plans can’t be used at enterprise. Enterprise pays expensive api rates
From the enterprise admin of Copilot, Cursor & Claude…enterprise controls.
MicrosoftGitHub understands what companies need to run their products securely and successfully (at scale™)./edit to add. Our Copilot bill went down due to change in billing type, mostly due to organization pooling of credits.
That just sounds like bad management, and a huge opportunity to cut costs by switching to Linux and something like cursor or other alternatives.
I don’t think you understood my sentence, we (the company) already offer plenty of choices. In fact we offer OpenAI (Codex, ChatGPT) as well as Gemini. I just admin Cursor, Copilot and Claude (and at one time we offered Windsurf, but that didn’t have enough users to rectify the costs/time spent administrating it). What I am trying to tell you is that the fairly new companies, OpenAI, Cursor, Anthropic/Claude absolutely suck at making their product usable by the enterprise while Gemini and Copilot completely understand enterprise needs/requirements.
We already have Linux laptops (with enterprise controls). /edit Apple machines are primarily used though. We even offer Windoze if you really need it! (Some folks do.)
That’s fair, as long as it’s recognized that it’s that way for now, and rapidly changing.
You mean you guys don’t rotate between 10 free accounts and use their monthly quotas?
That many free accounts is against their user agreement though hehe
What are they gonna do, close your accounts and make you sign up for 10 more? lol
Limit it per household/IP and block VPNs. You people really have no imagination.
This is happening with GitHub because they resell other models, right? Including ChatGPT and Claude.
This hasn’t happened with the models directly on their platforms yet for the plans that still offer flat rates. Right?
At my company, all of us devs have GitHub copilot enterprise licenses, which was $39/month for basically unlimited use on most models. The new pricing is the same price except we get 3900 copilot credits a month. 1 credit is 1000 tokens. There is nothing concrete about how many tokens are used, but they estimate powerful models like Claude will use about 10 credits per transaction. The more basic models are less consumption, but still count. Code completion is still unlimited. It’s going to get really expensive really fast for companies balls deep in it.
For token context, we had someone come in and give us an hour long vibe coding lesson with Claude code.he created a basic 600 line app in that hour and consumed 40000 tokens. I have UI tests with twice as many lines of code. It’s going to be a blood bath, and I’m here for it.
You can buy energy, which you can use to get crystals or exchange for gems. Using the service costs karma, which you get from coins. But don’t worry, you can get coins easily just by using the service, or buy them with gems and crystals.
This whole token, credit, etc thing is modelled off of gatcha game mechanics, and designed to make you forget the actual cost.
It’s crazy that it was ever set up this way in this context.
The direct to end user flat rates make more sense. With their usage windows, and stuff like that. But even those will change someday, too, I bet. But I think this nuance is lost right now in all the joy and schadenfreude sourced from Microsoft fucking this up so badly.
Yesterday I had a meeting where I explained this to my team. Nobody can use Opus anymore and shoot for under 5% of your monthly budget per work day. They better pray that I am the first one to run out of credits, I won’t buy more until that. From today alone it looks grim, they changed something in the harness too, making it burn credits. I would say we use it very lightly and we are in trouble. It is going to be a blood bath. Does anyone know if the JetBrains plan is more token efficient? I am also looking at just paying API cost for a cheap model on openrouter.
We don’t have any power users so I think we’re going to be fine for now. I do believe that by default all of your company credits are pooled, so one person could use everyone’s in theory. I thought I read that can be changed.
I’ve asked 3 questions to Claude today because I was struggling with an error. I’m curious to see how many credits I used to end up not using any of the generated code.
I change it to not be pooled. I am not going to used them conservatively for someone else to blow the entire plan on me.
Did you find a way to see how many of those credits you already used? To me it looks like they have completely removed access to that information from individual devs and only admins can manually generate reports containing it.
I haven’t. They just went live on the new model yesterday. When I read the release, it said there will be tools to monitor all of that. Whether that means they exist and I don’t know or they’re going to release them late and half finished in typical MS fashion is unknown to me.
I had talks with my lead and manager so they are aware. I have strong doubts that has been related to the people that need to monitor that. The company has been trying to push AI adoption for the past year, and we all hate it, so I’m not going to go out of my way to stop them from stepping on their own dicks. Hopefully a stacked unexpected bill would convince stakeholders to sit back and listen to those of us in the know. It won’t, but I’ve got my popcorn ready.
Oops. Now that users are being to made to pay something closer to the true cost of AI inference, no one will like using it anymore. Could this be what ultimately sets off the bubble?
Users are paying way more than the cost of inference. Look up inference prices of high end open weight models vs claude or gpt. Cheaper by an order of magnitude.
It’s the constant training of new models that’s losing them money. New version is out every month.
To be clear, We don’t really know the architecture or model size of claude or chatgpt. If opus is a 1 trillion parameter dense model, then yes of course it’s going to be way more expensive to run than deepseak which is an 862 billion parameter MoE model. American models have focused on being the best regardless of cost where chinese model makers were forced to focus on efficiency because of lack of access to chips. Which will probably give them an advantage as the free investor money runs out for the american companies.
The constant training is simply how AI works and will continue to work chinese or american. That cost will never go away. Anytime you need your model to learn new skills or gain new knowledge, it needs to be trained.
To clarify, AI companies charging the cost that would make the inference profitable for them, against the operating costs and financing costs on new capital expenditures (new data centres, new compute and new model training*), is more than what most people appear to be willing to pay. That cost is indeed more than just the cost of inference incurred by the AI company.
*(I’m being generous and including model training as capex for the sake of argument, even if I personally think to continue the hypetrain, continuous model improvements are core to AI companies’ operation.)
Is the answer don’t use copilot, or am I just talkin’ crazy?
Delightful tears. Thanks.
The 2 remaining devs using Copilot will leave it.
It’s rare to see such a clear example of first-mover disadvantage as GitHub Copilot.Honestly, only Tesla comes to mind. Maybe bleeding edge tech frontier is the one place where first movers RND cost is heavy enough to make it irrecoverable.
It’s interesting to contrast this w/ the field of medicine; RND cost can be recovered by squeezing the folks in the domestic market.
The company I work for had a pilot project for a while but recently they opened up for all users to order Github Copilot. Of course it is Github Copilot since everything else at the company is from Microsoft.
"For basically nothing’… If it’s basically nothing then use your damn brain to do it.
Best I can do is spin up 20 agents to barely get some basic functionality written.
I’ve been saving this video for this particular moment
Another thing that greatly improves productivity is not relying solely on AI to do your job for you.
Good devs don’t rely on it to do their job, but use it to do their job better and more productively.
As a dev, copilot is amazing.
Well good luck with the next few years of sprint planning. Shouldn’t be long before the climate impacts really catches up to us.
It certainly saves time because of course plagiarism saves time. Does it mean people are doing their job better or more productively? Depends on how carefully they review that AI output, but I have a feeling that people aren’t always so vigilant. That’s where the technical debt creeps in.
My friend who uses Claude daily says it saves him a hell of a lot of time doing all the routine work, which he verifies. It’s not AI’s fault if devs are sloppy, or if non-devs think they can type “Create software” and go get coffee. I still remember some people around 1990 who similarly misunderstood object-oriented programming - like it meant drawing a box labeled “DoAccounting” and magic would happen.
What I want to know, is that are they just charging closer to the real cost or the actual real cost.
Chances are they would want to slowly increase the price à la boiling frog method.
And once that happens, then they have to increase it again to make profit, AND that has to measure up against regular ways of making money so it can’t just be barely profitable.
It’s a long road ahead for them
Single prompt takes 25% of the monthly usage in like 30 seconds even though I have Pro haha
Anyone planning to make a profit by billing this as metered is dreaming when agents suck down 40k tokens in a single prompt. I did the math, at like 0.2 cents a token that would still be 8k. No one is paying 8k for a single prompt. At 0.015, thats still 600 dollars. Its never going to make sense.
You might want to recheck your math.
.2 cents is $0.002
40000 * .002 = $80
Yeah I guess I did twenty cents, I realized this earlier when I paid for shipping at work. 80 dollars is still a lot for a single session. I think I’ve easily seen 100k+ sessions too, but Idk how many of those were just people burning tokens.
Edit: two cents is 0.02, 800 dollars. Just quadruple checked.
80 bucks for a session could be a lot, but if it saves you 4 hours of work, I’d say it’s worth it.
OTOH, I’ve seen some articles claim that using AI tools actually slows you down. In that case, it’s a giant boondoggle.
In the frame of a company, yes. For individuals, not really imo. What’s going to happen is the big models will probably move to being only accessible by corporations and rich while the public moves more towards open weights and self host able options.
True, I’m very much looking at it in terms of work tooling. I don’t have much use for the ELIZA-with-delusions-of-grandeur models in my personal life.
Two cents is $0.02.
Yes, and?
Point two cents is not the same as two cents.
Codeberg accounts incoming!
I don’t think they’re moving from GitHub, just GitHub CoPilot.
they should probably move away from plagiarismhub
That’s a non-existent distinction. GitHub is under Microsoft’s AI division.
https://www.techspot.com/news/109040-microsoft-ai-push-tightens-grip-github-after-ceo.html
