I am typically an AI skeptic, but...
An AI skeptic, who extensively uses LLMs for work, is reconsidering their previous economic critiques, which aligned with Ed Zitron's revenue concerns. Despite historical skepticism about the numbers, they now question why more people wouldn't pay $10 or $20 a month for LLM technology, finding it infinitely more useful than services like Netflix. They observe a clear, common, and consistent use case beyond programming, wondering why current adoption numbers don't reflect this perceived value.
时间与来源
时间显示为 UTC
显示时区：UTC
本地时区尚不可用，暂时显示 UTC。
发布当时偏移：UTC+02026年10月30日 01:41 UTC
收录当时偏移：UTC+02026年10月30日 04:01 UTC
- 发布
- 2026年10月30日 01:41
- 收录
- 2026年10月30日 04:01
- 来源类型
- 开发者社区
- 档位
- 社区
- 信源状态
- 正常
档位是按信源手工设定的编辑判断，不是逐条打分。
I use LLMs extensively for my work and did so years before that was mainstream. Usually my critique of LLMs is the economics; I think the technology itself is quite useful. My positions have historically been in the general area of Ed Zitron's — where's the revenue? But lately that's starting to shift.
Today, an average day, I used ChatGPT to estimate the calories in a meal, determine how I could add a few hundred more while improving its nutritional quality, all without straying from the meal's traditional cuisine. I needed help writing, recording and stitching together a series of videos for work, something I'm not particularly comfortable with, and was able to use the LLM to coach me through the process, with ease and in very short order. I then had it investigate a number of local retreat centres to see what meditation events were available over my winter break, accounting for all of my previously scheduled obligations. It found opportunities that didn't surface through my own Google searches, one of which I've now booked. This obviously saved me huge amounts of time and let me query the internet in a creative way, without having to trawl through hours' worth of websites, leaving me, a grown human, to verify and implement the results.
Once again, I have for a long time been genuinely skeptical of the numbers, but it isn't obvious to me that at $10 or $20 a month most people wouldn't want this technology. People pay that much for Netflix. This is infinitely more useful to me than Netflix, more life-affirming, more creative, more empowering. I'm starting to see that this is a technology with a very obvious, fairly common, and reasonably consistent use case, beyond programming. I do wonder why the current numbers don't reflect that, whatever about projected use — and perhaps I've entered into some kind of mental bubble here — but why doesn't everybody want to use this? And to bring it back to the numbers, hypothetically, why wouldn't most people pay for the services I'm describing?