I love how corpos can just change the rules at will.
Edit: New prices:
https://docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing
And if you look at the old pricing structure, some of the models are increasing by 27x
I mean, if AI gets too expensive, companies can always hire juniors to replace them 🤣
When the junior devs get too expensive they can outsource all of their software development to Bangladesh
When offshore-sourced code gets too shitty they can hire some senior engineers to rebuild it in a way that’s compatible with the rest of their ecosystem.
Which brings the model back to AI. All businesses have accomplished was basically add an extra step in actually forcing themselves to pay for what they actually need every so often. Instead of Sr>jr>offshore>sr. they now just added in AI to the loop.
Not that charlatans care, as long as they get theirs in the interim.
Users on annual Pro or Pro+ plans will remain on their existing plan with premium request-based pricing until their plan expires, however, model multipliers will increase on June 1 (see table).
holy shit, 9x the previous cost. which was already not great. I was on the fence about cancelling it, but thanks for making up my mind, MS
That’s been their business model for a while now. “Here’s something you also didn’t ask for”
Which is a crime, by the way, when you sell it together with a product you hold a monopoly for: https://en.wikipedia.org/wiki/Tying_(commerce)
I love that their email made it sound like it was a wash, like “we’re just changing our billing model, you’ll get credits now, samesies!” but then this pricing chart buried 3-levels down from the announcement lays out just how much less you’re going to get for the same price.
I wonder about all the startups who were bragging about their $10k/month AI coding bills being the best money they ever spent. When this new pricing kicks in and pushes it to $40k/month right around the time all the vibe-coded shit blows up their codebase, I wonder if they’ll still be so happy with their choices.
I interviewed for a place a while back and started asking about quality and velocity and how they balance it with AI developer tools, he said something like “One of our developers closed 400 PRs last month” and I instantly knew it was definitely not the place for me.
I wonder about all the startups who were bragging about their $10k/month AI coding bills being the best money they ever spent.
WTF did they think was going to happen?
So, did they use AI tools to type “LGTM” 400 times or nah?
But yeah, I also find that frustrating. Management just looks at terrible metrics like PRs closed or lines of code produced.
It’s not even novel that you can produce terrible code very quickly. Decades ago, our industry learned that it isn’t worth it, because you suffer for it later. Now the game is altered slightly and management demands that we throw all these learnings out the window.
They should really describe this as you’re on the same plan, but your plan gives you 80-88% less use.
These tech companies are a bunch of criminals. I fucking hate this industry.
You’re looking at Claude, I don’t See Copilot
Edit: ignore me, I finally reviewed the article and is through GH, not actual MS 365 page.
copilot isnt a model. its a front end that lets you pick a model. that table shows usage costs.
Gonna be hilarious when the people who haven’t been paying attention realize that they just replaced workers with shit that doesn’t work AND actually costs more.
Yep, I’ve been telling anyone who’ll listen that if you really want to drop juniors and give tools to seniors, then you have to pay the monthly cost (whatever it will be) and you have to be ready to foot the big bill in 5 years when your seniors (with no candidate replacements) say they’ll take a 50% raise or walk.
I could be really cruel right now, I could point out to people how this is chickens coming home to roost, how it’s always been grossly subsidized and underpriced, how the game was always getting users, companies, the country, the whole damn world hooked on these slop machines and then once they’re reliant on them to jack up the price hugely. I could be really mean about this, I could lord it over people, I could point out how they’ve been dupes this whole time and in the process traitors to all of humanity.
I could. i think I will.
New meme dropping: “telling you I told you so? Yes, I think I will”
People like you act like local models don’t exist. Anyone relying purely on a single cloud hosted model for their work is an idiot.
Anyone relying on AI to do their work is an idiot.
I have an acquaintance (not sure if qualifies as a friend) who made news for getting fired for, when they were ill, asking an AI to get a quote out of an article for them, which it just out-and-out made up and got them fired. They aren’t stupid, but maybe a bit too trusting. The world lambasted them, which is a huge shame because they do really good work. One strike, you’re out.
So I am not unsympathetic to people who get, in essence, betrayed by AI. (not literally so, because AI is not a person) But yes I think people should be very careful, more careful they they’re being now certainly.
They also should be ready for if it all falls apart, as training costs increase and becomes harder. Already websites like IMDB are putting CAPTCHAs on their site just to browse it. It used to be an accommodating site to scripts that used it to gather information automatically, but those days seem to be over now. Expect to see that more and more, and in the process, the web becomes that much more annoying to use for plain old human readers too.
So, if you subscribe to the $10/ month plan you get $10 worth of tokens to use within a month, and if you subscribe to the $49/month plan you get $49 worth of tokens to use within a month.
At this point why subscribe? Just pay directly the tokens so if you use less you pay less.
To me it doesn’t make sense to subscribe at $49 and maybe use $30 of tokens.
People were subscribing at the $49 plan because they were using $1000 in tokens worth
How else are they going to begin to recap their billions and billions of debt? Someone has to pay for all those data centres, all that hardware, all that power, etc etc etc. It will be the companies that have come to rely on AI.
Sure, for now, AI is a lot cheaper than an intern, but it doesn’t become an expert like a human does. And Amazon used to be cheaper than other retailers right up until they had achieved vast market share.
This cannot be the last 10x price multiplier they pull. Not even close. Firstly they’re way, way, way off from recouping their costs, and secondly, they’re still way, way off market value for an incompetent human intern who isn’t learning much.
Uber didn’t enter the market to open up taxis to new drivers and bring down prices, that was marketing. They entered the market to take a cut out of ever taxi fare in the world, and drive up prices at peak times to many times the agreed fares, especially in regulated areas.
Similarly, AI didn’t enter the coding market to drive down prices and enable greater access for folk to generate code. They entered the coding market to receive the wages of programmers and drive up prices in in-demand fields. They are not unaware of how much companies pay devs. Why else would they have spent all those billions in advance? Where is the payback coming from?
Just as open weight models are getting good. Qwen 3.6 27B just dropped with claimed performance approaching Opus 4.6, but it can run on a Mac with a M-series SoC. I tested it out today on a M4 Pro with Ollama and Cline and was impressed with its reasoning, but it was slow. Going to try with llama.cpp tomorrow and mess around tweaking it for speed.
https://ai.rs/ai-developer/qwen-3-6-27b-local-coding-model
AI coding agents are useful, but it’s time for the cloud-based models to chill out so we can get cheap RAM again to run our shit locally.
It’s almost like buying all the RAM so most people can only afford subscription services is the point.
https://www.reddit.com/r/GithubCopilot/ right now is entertaining
The tears are delicious
Getting de-skilled is starting to look like a very expensive gamble and this is just the first price hike, expect more to come. And expect them to criminalize open source models as well with some national security concerns or something.
Interesting.
Inline completions are genuinely useful, I’m mostly replacing them with local models though. They are slower but free as in beer (once you pay the hardware cost).
I tried AI auto-complete for maybe 6 months. I found that mistakes and hard to find bugs increased, actually lowering my productivity. I find that actually just typing everything (well, with traditional LSP) gives me more time to think about my code, the architecture, etc while I’m typing, leading to better results. “Slow is smooth, smooth is fast.”
I’m interested in setting it up, are you using vs code? Which extension or editor?
I’m using vim with minuet-ai, and it plugs the AI suggestion into my completion module. I found the Copilot style virtual text interfaces all janky.
So, not free. Just capital expense instead of operational.
Sure, but hopefully small code completion (2-4b range) models can run locally on a lot of things. They’re just less good.
Claude is cutting back on usage policy for pro users…
GitHub CoPilot is now doing it too…
It’s not hard to see the future, AI companies bought up all the RAM creating a shortage which raises prices for all of us across the board.
Now they’re going to throttle and limit access to it behind a paywall per transaction. The future is so stupid. Can we just go back to dial up internet and IRC? It was a much simpler time.
You could also just not use these services. You don’t need them. They were a stupid idea to begin with.
You could also just not use these services.
I do need RAM, though.
Fortunately, I got pretty well stocked up before this shit, but the hardware I’ve got now will fail eventually.
It’s why I only use the free versions of these lol
Man, they couldn’t have communicated this more confusingly, if they tried.
That was intentional. They were trying to word the announcement to not make it sound like you’re now getting 1/5-1/9 as much AI for the same price.
Yeah, I imagine that they did try. But it’s not just the intentionally misleading announcement post, they also have 5(?) different subscription tiers, which get different changes from this. And one of the subscription tiers is actually called “Pro+”, so that does not mean “Pro and more expensive tiers” like I wondered. And they have this ridiculous intermediate currency to make things even more confusing.
Their offering itself is overly complex and confusing…
The whole “AI Credit” thing did strike me as odd to even introduce. They mention a few times that 1 AI Credit = $0.01, but then do the whole pricing table in $/1M tokens.
The only place I saw them even use “AI Credits” as a unit was to say the $10/month plan includes 1000 AI Credits. Why even introduce a whole new unit if you only use it to say your $10 monthly plan includes $10 of usage?
I suspect the answer is so that they can later muddy the waters by changing the number of included credits to be less than you’re paying monthly without directly saying “you’re now paying $20/month for $10 of usage”
I guess it is coming from the same people who came up with the world’s most inscrutable billing scheme for compute…
I mean, even then, they could increase the price per token, if they want to hand out fewer tokens for the price paid.
They could make this work like a prepaid SIM card, where you charge it with e.g. $10 and then you can use it until the $10 are used up.
Instead, they make it work like in-game currencies in scammy free-to-play games. Except that they didn’t choose a confusing conversion rate, for some reason…
I literally just signed up for the $10/month plan… Guess ill just run my own git server.
Git? Forgejo is really nice.
Im just familiar with git.
Forgejo is a git server.
UI looks very similar to older GitHub before it got so “modern” in a bad way.
Gotcha, ill give it a go. Thanks
What does having a git server have to do with copilot subscriptions?
Moving local to not support githib, I run Open Claw local too
