Earlier quoted context omitted.
Don't worry, they'll just lobby to ban Chinese models instead to keep their token revenues high. > Compounding the problem, labs in China often release dual-use capable models as open-weight. Once a model is open-weight, safeguards that do exist can be removed, making the model available to any state or non-state actor to use for malicious purposes, including the cyber and CBRN misuse those safeguards were built to p…
> Once a model is open-weight, safeguards that do exist can be removed Safeguards trained into the model (ie exist in the weights) can’t be removed.
Uber's $1,500/month AI limit is a useful signal for AI tool pricing
691–700 of 819 posts
Re: Uber's $1,500/month AI limit is a useful signal for AI tool pricing
#692> I noted that my own token usage comes to about $1,000/month against each of Anthropic and OpenAI - which currently costs me just $100 per provider thanks to their generous subsidized plans for individual subscribers. Do we know that AI providers are going to keep these per-token prices, or eventually lower them because of competition from China? Many lower-budget individuals are now moving to China open weight mode…
Re: Uber's $1,500/month AI limit is a useful signal for AI tool pricing
#693Re: Uber's $1,500/month AI limit is a useful signal for AI tool pricing
#694Earlier quoted context omitted.
And Chinese electricity prices are some of the lowest
Don't know why people keep parroting this, this is incorrect. Chinese electricity prices are equal or slightly cheaper then most of North America. But significant pockets such as those around the Quebec or other hydro plants are significantly cheaper then Chinese power pricing. Not only that, China may subsidize AI, but so does the US.
Re: Uber's $1,500/month AI limit is a useful signal for AI tool pricing
#695Re: Uber's $1,500/month AI limit is a useful signal for AI tool pricing
#696Earlier quoted context omitted.
They don't have display-out. You'd have to send back the screen data over pcie to the motherboard for display.
Not exactly a problem for cloud gaming.
Re: Uber's $1,500/month AI limit is a useful signal for AI tool pricing
#697Earlier quoted context omitted.
Genuine question: what would make me a "paid shill"? Who do you think would be paying me, and what would they expect in return?
Your unwavering praise of LLMs' performance which does not match anyone's reality? OpenAI or Anthropic would be paying you, like they pay bot farms and other influencers, and they would expect marketing in return, which you provide in boatloads. Your job is to be an influencer, I'm not sure why anyone would be surprised that this is a possibility.
The reason so many people read my writing and find it useful is that they see me as a credible source of information: in a world full of clickbait and misinformation, I have a reputation for providing an independent voice that occupies that rare middle ground between "AI will kill us all" doomerism and "AI will solve everything" hype.
Credibility is hard to earn and easy to squander. I've been blogging for 24 years now, which has helped me build credibility with a large array of people across many different interest areas.
The modern influencer business model is to grow an audience and then sell things to them, through partnerships and sponsored content. I refuse to do that, because it strikes directly at that credibility. The moment you say "I've partnered with X to tell you about product Y" you're no longer an independent voice.
Nilay Patel of the Verge (and the excellent Decoder podcast) refuses to read ads from sponsors himself, at significant financial cost to his publication. I've adopted the same policy - I will not let anyone else pay me to put words in my mouth, because it strikes directly at the credibility I value so much.
Until a few months ago the only money I made from my blog was an https://ethicalads.io banner which pulled in a few hundred dollars a month (more if I had a high traffic piece). It helped cover some of my hosting costs for my various projects.
That changed in February - https://simonwillison.net/2026/Feb/19/sponsorship/ - when I added a Troy Hunt-style sponsor banner to my site (no cookies, no JavaScript) - currently sold by an agency called Freeman & Forrest. Sponsored slots are sold on a weekly basis and get a mention in my email newsletter in addition to the blog banner.
I'm earning enough from those that I no longer feel the opportunity cost of not going and getting a proper Silicon Valley engineering job.
If I was a publication like the Verge I'd have a complete firewall between editorial and advertising. I don't have a team, but I've tried to replicate that as much as I can by having Freeman & Forrest sort out the sponsors while I stay hands off. I'll veto sponsors if I have to (no prediction markets etc) but thankfully that hasn't been necessary so far.
I maintain a disclosures section on my blog here: https://simonwillison.net/about/#disclosures - which was inspired by Molly White's: https://www.mollywhite.net/crypto-disclosures/
I'm currently considering extending that to more of an ethics statement like this one on the Verge: https://www.theverge.com/ethics-statement
The Verge policy I'm currently not fulfilling is "Our policy against receiving anything of value from companies we cover includes, but is not limited to, things like gifts, meals, discounted services, or paid trips and junkets. Vox Media and The Verge pay for all travel expenses to all events, including transportation, food, and hotels." - I've occasionally accepted flights, dinners, accommodation and some pretty absurd swag (Microsoft just gave me a jacket with my name stitched onto it as part of the GitHub Stars programme, and a bunch of gadgets in a pelican case) which didn't bother me so much when the blog was a side project, but I think I need to start refusing those kind of gifts.
The day after the jacket I wrote a piece about their new models - https://simonwillison.net/2026/Jun/2/microsofts-new-models/ - which I later had to update because I missed some crucial details. Was I subconsciously influenced by the freebies? I don't think so, but the whole point of "subconsciously" is you don't know for sure.
Re: Uber's $1,500/month AI limit is a useful signal for AI tool pricing
#698Earlier quoted context omitted.
Here's my AI misuse tag: https://simonwillison.net/tags/ai-misuse/ - 54 posts My ongoing coverage of AI ethical issues: https://simonwillison.net/tags/ai-ethics/ - 308 posts I've been the loudest voice about the fundamental insecurity of LLMs for several years: https://simonwillison.net/tags/prompt-injection/ - 150 posts In https://simonwillison.net/2025/Aug/25/agentic-browser-securi... I said "I strongly expect that…
Literally none of those articles are critizing LLMs, only use made of them by 3rd party actors outside of the providers. It really has nothing to do with LLMs themselves. The fact that you had to dig to August 2025 to find a single article that's actually a critic of something produced by the AI labs is just further proof.
Re: Uber's $1,500/month AI limit is a useful signal for AI tool pricing
#699Re: Uber's $1,500/month AI limit is a useful signal for AI tool pricing
#700Earlier quoted context omitted.
I used to work in datacenters, during spinning disk era we had technicians from vendors basically every couple of days to replace some broken part. When the massive switch to ssd happened instead of having them every couple of days it was 3 or 4 times per month. Despite no moving parts things broke anyway and, even if it doesn't break, the vendor can make you change the technology just by playing with maintenance cos…
My understanding is that a lot of AI data centers are still heavily relying on spinning HDDs, which is why seagate, western digital are selling more HDDs than ever before.
https://s24.q4cdn.com/101481333/files/doc_financials/2026/q3...
"Hard Drive exabyte shipments of 199EB, up 39% YoY, with ~90% shipped to data center customers"
"Data center revenue of $2.5B, up 55% YoY, driven by strengthening cloud and enterprise demand"
And an article: https://www.seagate.com/stories/articles/the-ai-era-doesnt-r...