Project 911reg MapExtractor

[aguy]

8 minutes ago, xdem said:

 

Ask me if I give a fuck to debate anyone in here before baiting me into a stupid debate like yours. Anthropic and OpenAI have already tightened usage limits and increasingly differentiated access to their most capable models not once but multiple times, without prior notice. Enterprise-scale usage can cost vastly more (20x per token) than a flat-rate consumer subscription, while those subscriptions don't reflect the true cost of unrestricted compute costing these companies billions.

Doesn't that tell you something? If not piss off and never quote me again.

 

Where did you learn that economics, same place you learned coding ? Can you even read a financial statement ? Let me educate you a bit:

 

Anthropics operating revenue is positive. What users pay - what it costs to run  (electricity gpu depreciation) is POSITIVE. What is subsidized is the cost of training new models, the RnD part which is huge. Which means that in order to compete in the AI race, they have to get investors money to cover huge research expenditures. 

In other words: For each dollar you PAY anthropic, they pocket about 40% as gross profit after you remove the costs of running the inference business.

 

In IT companies, RnD because it's human spending mainly, by accounting, is obligated to put it in the income statement as expense. In ALL other companies, this is CAPEX, and goes into the cash flow statement so the net income looks positive.

Do me a favor, if you ever get money on your hands, NEVER try to invest them yourself.

[rexen]

[xdem]

10 minutes ago, aguy said:

 

Where did you learn that economics, same place you learned coding ? Can you even read a financial statement ? Let me educate you a bit:

 

Anthropics operating revenue is positive. What users pay - what it costs to run  (electricity gpu depreciation) is POSITIVE. What is subsidized is the cost of training new models, the RnD part which is huge. Which means that in order to compete in the AI race, they have to get investors money to cover huge research expenditures. 

In other words: For each dollar you PAY anthropic, they pocket about 40% as gross profit after you remove the costs of running the inference business.

 

In IT companies, RnD because it's human spending mainly, by accounting, is obligated to put it in the income statement as expense. In ALL other companies, this is CAPEX, and goes into the cash flow statement so the net income looks positive.

Do me a favor, if you ever get money on your hands, NEVER try to invest them yourself.

 

You wrote an entire lecture on gross margins while completely missing my point.

I never argued that inference has negative unit economics. We're talking about the future concentration of compute, infrastructure, and access to frontier intelligence, not Anthropic's profitability today.

And since you're lecturing me on accounting, R&D isn't automatically CAPEX outside IT. Under US GAAP, it's generally expensed across industries.

The point is simple: technological progress doesn't guarantee accessibility. The hardware, capital, and infrastructure required to remain competitive are increasingly concentrated in fewer hands.

And forgive me for not blindly trusting a handful of corporations to keep frontier intelligence affordable once entire industries become dependent on them.

You trust technological progress to preserve accessibility. I don't trust centralized corporate power to preserve it for us.

That's our disagreement, not your accounting lesson, now fk off Im blocking you.

8 minutes ago, rexen said:

I dont trust Musk either at all, he just plays the other side to pass the negative narrative to the masses so they are ready to adapt peacefully since they have been warned already.

[aguy]

1 minute ago, xdem said:

That's our disagreement, not your accounting lesson, now fk off Im blocking you.

Oh no ! Don't block me, I'll appeal to the European Court of Human Rights !

[eMommy]

2 hours ago, aguy said:

If I recall how many collegues have told me that "next year" Anthropic will double the prices in 2024, 2025, 2026 ... Shortages can spike the prices, but as technology becomes better, frontier technology becomes comodity. The gap between open weight models and private frontier models is getting narrower month after month. 

 

Progress is here to stay. It's very rare for humanity to regress in technology. I'm not saying it can't happen, but it has happened 2-3 times in the history of our civilization.

 

AI agentic coding is here to stay. There's no ride to enjoy, we are going through a shift in the way we produce software. That's all. 

This is a very European dogma. To reject modernity and progress. Regression is the European form of progression. Hence why in 2026 there are no air conditioners in major countries like Spain France. Hubris of their ancestor's accomplishments trickled down to the modern person (especially itlians and greeks) where they still believe that their situation is that of 2000 bc.

 

While I do agree it is definitely important to understand your craft; the reality is as stated. The bots are here to stay and it's as revolutionary as the invention of the internet or the computer. Purists can stick to the typewriter while the keyboard is invented or snail mail containing hand written letters meanwhile the world carries on with hyperlinks. Either harness the power of superintelligence or get left behind. No matter the cost of tokens imo it is still worth it rather than dealing with human assets. A 'decent' single developer has now became an entire dev team in the l2 space. No complaints, no hours, no breaks (except when you run out of tokens. If this is a problem then just buy another sub???) and any mistake can be corrected ASAP. As long as the prompter has a "good enough" understanding of the context and is able to articulate to the bot. Any moron that talks to claude can do it. There is nothing special, you and I can both do the same thing now. AI levels the playing field. The difference is the macro understanding and how to fine tune it. Getting the feature or fix "working" is now no longer the problem but the perhaps the last 20% of making it user friendly and ergonomic which still requires a good eye and experience.

 

Pre-AI I have spent maybe 12k euro on brazilian scammer gabriel charging 40$ an hour, and then when critical issues arise from their code, the response is just "o sorry im at lunch" so now there really is not that much worry? Or waiting for someone to work on a major code refactor for over a year after buying them a laptop only to say that they are busy herding 5000 sheep and can no longer work. I also spent over 20k on reverse engineering the client and creating extenders which now can be solved with the same solution. The quality is maybe 70-80% of an expert, but the function is without a doubt 90-95%. However I have now adapted to the forward thinking workflow. I complain to opus, go get a drinky drink and come back and try the fix. I can also be talking to the bot on my phone and having it work on my pc while i do other stuf. All for the measly price of about 520$ a month (Claude 200, Claude 100, GPT 200, Gemini 20). What used to be impossible is now just a matter or tokens and persistence.

 

Also posting a tweet from Elon is the same as posting a quote from Mobius, showing the user's inability to think critically and how they are prone to idolize the loudest in the room.

[xdem]

22 minutes ago, eMommy said:

This is a very European dogma. To reject modernity and progress. Regression is the European form of progression. Hence why in 2026 there are no air conditioners in major countries like Spain France. Hubris of their ancestor's accomplishments trickled down to the modern person (especially itlians and greeks) where they still believe that their situation is that of 2000 bc.

 

While I do agree it is definitely important to understand your craft; the reality is as stated. The bots are here to stay and it's as revolutionary as the invention of the internet or the computer. Purists can stick to the typewriter while the keyboard is invented or snail mail containing hand written letters meanwhile the world carries on with hyperlinks. Either harness the power of superintelligence or get left behind. No matter the cost of tokens imo it is still worth it rather than dealing with human assets. A 'decent' single developer has now became an entire dev team in the l2 space. No complaints, no hours, no breaks (except when you run out of tokens. If this is a problem then just buy another sub???) and any mistake can be corrected ASAP. As long as the prompter has a "good enough" understanding of the context and is able to articulate to the bot. Any moron that talks to claude can do it. There is nothing special, you and I can both do the same thing now. AI levels the playing field. The difference is the macro understanding and how to fine tune it. Getting the feature or fix "working" is now no longer the problem but the perhaps the last 20% of making it user friendly and ergonomic which still requires a good eye and experience.

 

Pre-AI I have spent maybe 12k euro on brazilian scammer gabriel charging 40$ an hour, and then when critical issues arise from their code, the response is just "o sorry im at lunch" so now there really is not that much worry? Or waiting for someone to work on a major code refactor for over a year after buying them a laptop only to say that they are busy herding 5000 sheep and can no longer work. I also spent over 20k on reverse engineering the client and creating extenders which now can be solved with the same solution. The quality is maybe 70-80% of an expert, but the function is without a doubt 90-95%. However I have now adapted to the forward thinking workflow. I complain to opus, go get a drinky drink and come back and try the fix. I can also be talking to the bot on my phone and having it work on my pc while i do other stuf. All for the measly price of about 520$ a month (Claude 200, Claude 100, GPT 200, Gemini 20). What used to be impossible is now just a matter or tokens and persistence.

 

Also posting a tweet from Elon is the same as posting a quote from Mobius, showing the user's inability to think critically and how they are prone to idolize the loudest in the room.

 

You spent several paragraphs explaining how useful AI is today, something nobody here disputed, not that if anyone did would add value anyways.

I'm using AI myself, extensively. I'm not arguing against progress, I'm questioning the long-term concentration of the infrastructure that makes that progress accessible.

Your entire argument boils down to "AI is great today, just buy another subscription tomorrow." That's not an answer to concerns about future dependency, market concentration, or affordability.

And there's something ironic about calling AI democratizing while casually suggesting everyone should just spend another $500 a month if they hit their limits. Technological progress and centralized control aren't mutually exclusive. That's the distinction you all keep missing.

As for the European history lecture, I'm not sure what air conditioning has to do with semiconductor supply chains or market concentration that is driven by corporate greed and macropolitics.