Steinberger appeared unconcerned about the bill — easy enough when you’re not paying for it — describing the spending as research into how software development would change if token costs weren’t a constraint. Everything his team builds, he noted, remains open source.
Follow Tom's Hardware on Google News , or add us as a preferred source , to get our latest news, analysis, & reviews in your feeds.
Luke James is a freelance writer and journalist.\u00a0 Although his background is in legal, he has a personal interest in all things tech, especially hardware and microelectronics, and anything regulatory.\u00a0 ","collapsible":{"enabled":true,"maxHeight":250,"readMoreText":"Read more","readLessText":"Read less"}}), "https://slice.vanilla.futurecdn.net/13-4-23/js/authorBio.js"); } else { console.error('%c FTE ','background: #9306F9; color: #ffffff','no lazy slice hydration function available'); } Luke James Social Links Navigation Contributor Luke James is a freelance writer and journalist. Although his background is in legal, he has a personal interest in all things tech, especially hardware and microelectronics, and anything regulatory.
usertests "Burned" but actually perfect when the company is pushing its employees to use more free tokens, as we've seen stories about for weeks. Reply
vanadiel007 He should us AI to analyze his costs and see if he could cut out some AI agents working for him, or combine them into a more efficient group to reduce cost. Reply
timsSOFTWARE He works for OpenAI. I'm sure they would like to see others copy his technique. Reply
badmanchillin At $1.3M/month ($15.6M/year) one can hire 70 senior engineers working full time. There's no way 3 engineers with unlimited access to Codex will outperform 70 equally competent engineers who also can use AI as reasonable people. This is just completely stupid. Reply
sygreenblum Yeah, I'm gonna have to agree with this but add that those 3 engineers may be running 100 agents (building, watching for deviations and hallucinations) but who's doing the code review and testing? Nobody is going into coding to be a tester. It's like becoming a doctor to be a brain surgeon but they're only hiring proctologists. Reply
AI_Kills The mortality cost of carbon (MCC), a scientifically derived and broadly accepted calculation of the cost of human life from carbon emissions, is 4,434 metric tons per human life. Carbon intensive queries like those done w ChatGPT 5.5 emit an estimated .37 grams/1k tokens or 11.73 billion tokens per human life expectancy. His GPT use this month cost the equivalent of 5 human lives in shortened life expectancy from the associated carbon emissions. This isn’t an accomplishment, it’s a felony. Reply
usertests AI_Kills said: The mortality cost of carbon (MCC), a scientifically derived and broadly accepted calculation of the cost of human life from carbon emissions, is 4,434 metric tons per human life. Carbon intensive queries like those done w ChatGPT 5.5 emit an estimated .37 grams/1k tokens or 11.73 billion tokens per human life expectancy. His GPT use this month cost the equivalent of 5 human lives in shortened life expectancy from the associated carbon emissions. This isn’t an accomplishment, it’s a felony. My agentic AI is a hit man? Hell yeah, gonna smoke out some orphans. Reply
ejolson badmanchillin said: At $1.3M/month ($15.6M/year) one can hire 70 senior engineers working full time. There's no way 3 engineers with unlimited access to Codex will outperform 70 equally competent engineers who also can use AI as reasonable people. This is just completely stupid. That's right. So how many humans would be needed to accomplish the same task as the three engineers with unlimited codex? Figure out their pay and work backwards to determine how much value all those AI tokens are worth. I think it's possible the three engineers would accomplish more without so many tokens. While I'm admittedly a rank amateur, my experience is that AI works well as a quick index into the documentation and simple code examples, but it quickly becomes inefficient for AI to write finished code. This could be domain dependent or a result of me not holding it right. Hopefully AI will become cheaper as hardware optimised for inference becomes more widely available. Reply
American2021 ChatGPT and I argued so much that ChatGPT filed for a divorce. She's dumber than dirt. Fortunately, I kept the house but am on the hook for the duration of the annual membership. I went out to celebrate and met Miss Gemini. We are engaged now. Reply
JohnyFin Yes….fire people, use AI agents to create bad codes, release them to achieve catastrophic scores. People make mistakes but AI make unpredictable mistakes. Besides, these AI agents still codes from all sources, no doubt. Creepy time for humanity Reply
Key considerations
- Investor positioning can change fast
- Volatility remains possible near catalysts
- Macro rates and liquidity can dominate flows
Reference reading
- https://www.tomshardware.com/tech-industry/artificial-intelligence/SPONSORED_LINK_URL
- https://www.tomshardware.com/tech-industry/artificial-intelligence/openclaw-creator-burns-through-1-3-million-in-openai-api-tokens-in-a-single-month#main
- https://www.tomshardware.com/subscription
- [Daily Due Diligence] NVDA NVDA
- Tag, You’re It: GeForce NOW Levels Up Game Discovery With Xbox Game Pass and Ubisoft+ Labels
- Samsung meeting transcripts show memory workers offered incredible 607% bonus worth $477,000, while logic chip staff get as little as 50% — union says misbalanc
- OpenAI’s New GPT-5.5 Powers Codex on NVIDIA Infrastructure — and NVIDIA Is Already Putting It to Work
- It’s Gonna Be May: 16 Games Hit the Cloud This Month, With More NVIDIA GeForce RTX 5080 Power
Informational only. No financial advice. Do your own research.