Top all-time · #1883
Martin Alderson
Personal blog · martinalderson.comaibusiness
Martin Alderson's personal blog at martinalderson.com covers ai and business. Since 2025, 41 different Hacker News readers have linked it 125 times, most often “Are OpenAI and Anthropic losing money on inference?”. It is still writing: the last post was on 2026-10-29, and you can follow it by feed.
Most-linked posts?
- Are OpenAI and Anthropic losing money on inference?9 people · 11 mentions · first linked 2025-08-28
- No, it doesn't cost Anthropic $5k per Claude Code user6 people · 7 mentions · first linked 2026-03-09
- Which programming languages are most token-efficient?6 people · 7 mentions · first linked 2026-01-09
- Open weights are quietly closing up – and that's a problem5 people · 5 mentions · first linked 2026-05-06
- Which web frameworks are most token-efficient for AI agents?4 people · 4 mentions · first linked 2026-02-23
- The Coming AI Compute Crunch4 people · 4 mentions · first linked 2026-01-10
- I'm (mostly) picking models on speed now, not intelligence3 people · 3 mentions · first linked 2026-08-02
- Is the AI Compute Crunch Here?2 people · 3 mentions · first linked 2026-03-08
- Are we repeating the telecoms crash with AI datacenters?2 people · 3 mentions · first linked 2025-10-26
- The first known runaway AI agent – or a bad marketing stunt?2 people · 3 mentions · first linked 2026-07-23
What Hacker News says?
https://martinalderson.com/posts/are-openai-and-anthropic-re... All the labs are going hard on training and new GPUs. If we ever level off, they probably will be immensely profitable. Inference is cheap, training is expensive.
Comment linking /posts/are-openai-and-anthropic-really-losing-money-on-inference/
on “You should write an agent”Submitted Are OpenAI and Anthropic losing money on inference? as a story · 515 points
HN discussionIn short, the others have a huge margin if you ignore training costs. See https://martinalderson.com/posts/are-openai-and-anthropic-re... for details.
Comment linking /posts/are-openai-and-anthropic-really-losing-money-on-inference/
on “Kimi K2 Thinking, a SOTA open-source trillion-parameter reasoning model”Submitted Has the cost of building software dropped 90%? as a story · 421 points
HN discussionI disagree on that and covered a lot of it in this blog (sorry for the plug!) https://martinalderson.com/posts/are-we-really-repeating-the...
Comment linking /posts/are-we-really-repeating-the-telecoms-crash-with-ai-datacenters/
on “IBM CEO says there is 'no way' spending on AI data centers will pay off”Submitted AI agents are starting to eat SaaS as a story · 411 points
HN discussionChatGPT’s revenue means nothing if reports are to be believed that it loses money on each paying customer on just inference. It’s definitely not enough to support its training costs. Sam Altman: We're very profitable on inference. https://simonwillison.net/2025/Aug/17/sam-altman/#:~:text=Su... Independent analysis: Inference is very profitable. https://martinalderson.com/posts/are-openai-and-anthropic-re... https://…
Comment linking /posts/are-openai-and-anthropic-really-losing-money-on-inference/
on “Leak confirms OpenAI is preparing ads on ChatGPT for public roll out”Submitted Are we repeating the telecoms crash with AI datacenters? as a story · 241 points
HN discussionThose are effectively made up numbers, since they're given to him by an anonymous source we have no way of corroborating, and we can't even see the documents themselves, and it contradicts not just OpenAI's official numbers, but first principles analyses of what the economics of inference should be[1] and the inference profit reports of other companies, as well as just an analysis of the inference market would sugge…
Comment linking /posts/are-openai-and-anthropic-really-losing-money-on-inference/
on “Ilya Sutskever: We're moving from the age of scaling to the age of research”Submitted What happens when coding agents stop feeling like dialup? as a story · 137 points
HN discussionI don't think that's true anymore. I've seen a few benchmarks where svelte results in less tokens than React frameworks to complete the task. There was a time when LLMs were awful at svelte (especially v5) due to stale knowledge but that's not the case today. The knowledge cutoff isn't so bad and LLMs are much better at copying the patterns in your own codebase rather than relying on baked in knowledge. Edit: This i…
Comment linking /posts/which-web-frameworks-are-most-token-efficient-for-ai-agents/
on “SvelteKit 3”Submitted Has Mythos just broken the deal that kept the internet safe? as a story · 22 points
HN discussion> Nothing on infra costs, hardware throughput + capacity (accounting for hidden tokens) & depreciation That's because it's coming at things from the other end: since we can't be sure exactly what companies are doing, we're just going to look at the actual market incentives and pricing available and try to work backwards from there. And to be fair, it also cites, for instance, deepseek's paper where they talk about w…
Comment linking /posts/are-openai-and-anthropic-really-losing-money-on-inference/
on “How OpenAI uses complex and circular deals to fuel its multibillion-dollar rise”Submitted The AI margin collapse is gathering pace as a story · 7 points
HN discussionI wrote about this a couple of weeks ago. It's actually often the biggest cost and it tends to be hidden away on most platforms! https://martinalderson.com/posts/watch-out-for-cache-read-co... Btw I still haven't came across any decent model that is <$0.01/MTok cache costs apart from deepseek thru their official API (even with the price increases). Seems like a bit of an opportunity for someone to take - drop cache…
Comment linking /posts/watch-out-for-cache-read-costs/
on “Hy4 preview”Submitted Open weights are quietly closing up – and that's a problem as a story · 5 points
HN discussionRemarkable that it echoes, from a different angle, this post from just a few days ago on HN: https://martinalderson.com/posts/has-the-cost-of-software-ju... This last post was largely dismissed in the comments here on HN. Simon's experiment brings new ground for the argument.
Comment linking /posts/has-the-cost-of-software-just-dropped-90-percent/
on “I ported JustHTML from Python to JavaScript with Codex CLI and GPT-5.2 in hours”Submitted No, it doesn't cost Anthropic $5k per Claude Code user as a story · 4 points
HN discussion- nl · 2025-12-24
Lots of sources, and you can do the math yourself. Here's a few good ones: https://github.com/deepseek-ai/open-infra-index/blob/main/20... (suggests Deepseek is making 80% raw margin on inference) https://www.snellman.net/blog/archive/2025-06-02-llms-are-ch... https://martinalderson.com/posts/are-openai-and-anthropic-re... (there's a HN discussion of this where it was pointed out this overestimates the costs) https:…
Comment linking /posts/are-openai-and-anthropic-really-losing-money-on-inference/
on “Local AI is driving the biggest change in laptops in decades” Submitted Could Excel agents unlock $1T in economic value? as a story · 4 points
HN discussionInteresting! I wrote this approach up (more or less - extract design system -> make templates -> export) some time ago and I've found it unbelievably powerful: https://martinalderson.com/posts/how-to-make-great-looking-c... . I use it all day every day with Claude Code. I sometimes wonder past code if this has had the biggest impact on my day to day productivity, either having to make do with semi-bad looking report…
Comment linking /posts/how-to-make-great-looking-consistent-reports-with-claude-code-cowork-codex/
on “Claude Design”Submitted How the AI boom might end: 5 scenarios as a story · 3 points
HN discussionInteresting. I did finally find a use for IPv6 which I wrote up here: https://martinalderson.com/posts/i-finally-found-a-use-for-i... Tbh though the docker problems are very serious and extremely painful to work around. Everything works great apart from Docker which has so many issues - it does not handle IPv6 inbound but IPv4 out well at all (at least as far as I can tell!).
Comment linking /posts/i-finally-found-a-use-for-ipv6/
on “I spent a week without IPv4 (2023)”Submitted Have the frontier labs mixed up AI safety and security? as a story · 3 points
HN discussionThis video is unconvincing. The thesis of the video basically boils down to "AI companies are using electricity, driving up prices, and residential consumers are paying for it". However it neglects that most of the power usage is caused by the same residential users, who are getting something out of it. 80–90% of AI energy usage is estimated for inference (eg. generating responses on chatgpt), not training[1]. Moreo…
Comment linking /posts/are-openai-and-anthropic-really-losing-money-on-inference/
on “We Found the Hidden Cost of Data Centers. It's in Your Electric Bill [video]”Submitted What GLM-5.3 Flash running on Chinese hardware means as a story · 3 points
HN discussionYes agreed - I wrote this up a while back https://martinalderson.com/posts/whats-going-on-with-gemini/ My view then was they are optimising the models for inference ability on their own hardware AND use cases, which is often speed and time to first token. They've somehow seemed to end up with terrible compute shortages, which again is surprising given how good Google is at infra deployments AND have their own hardwa…
Comment linking /posts/whats-going-on-with-gemini/
on “Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber”Submitted The Summer of Open Weights as a story · 3 points
HN discussionSell me that bridge then, because I believe OpenAI's staff who say it was because inference costs went down: https://twitter.com/TheRealAdamG/status/193244032829380632 Generally I'm skeptical of the idea that any of the major providers are selling inference at a loss. Obviously they're losing money when you include the cost of research and training, but every indication I've seen is that they're not keen to sell $1…
Comment linking /posts/are-openai-and-anthropic-really-losing-money-on-inference/
on “Why Everybody Is Losing Money On AI”Submitted The first known runaway AI agent – or a bad marketing stunt? as a story · 3 points
HN discussion
