here on HN in 2013. I'd read "Amusing Ourselves to Death" so it particularly resonated with me at the time. I've watched the world get even more sucked in each passing year.
When I use it for writing I usually have an idea or thesis to explore already. I typically write it out stream of consciousness. The core idea is mine.
I'm not using LLMs to think for me, instead just edit the writing so it has proper grammar and the run ons are reduced.
Even better for something that has data. You can pull PDFs with encoded tables and extract the data into consistent units to compare. You can have it write code to do it and cross check the outputs. You can do that in a minute. Would take hours without an LLM.
My step mom was having debilitating pain. A year of going to doctors and no one was able to find a cause. I scanned her discharge paper work which had her prescriptions on it and gave it to Claude. It identified a prescription that had that exact side effect. They later confronted her primary care that concurred and took her off it.
A friend of mine's wife recently passed. They were chasing a suspected heart defect for over a year. She had been intermittently fainting. At about the year mark they decided to scope her digestive track. They found bleeding ulcers from cancer that was all over her body. I input her fainting symptoms into Claude and gastro impact was number two suspected after heart issues.
I have a few of other cases it's helped with. I'm not sure it could do worse than my own experience with the medical system. This is doubly true in places that lack any sort of medical care.
My mom had cancer and she was on regular, suppressive chemotherapy. I put her info into an AI and it correctly noted that her chemotherapy had stopped being effective 2 months prior based on factual lab reports. She was unaware of this. I was able to be her health advocate much more effectively by respectfully asking her oncologist targeted questions. He was already on top of it and was addressing the issue. Our conversation was respectful and, due to my educating myself, went up another level. Ultimately, it was a positive interaction. I was satisfied that he was indeed expert at his craft, and he was satisfied that we were aware of the uncertainty of the new treatment with a risk-based understanding of the viability of success. This was a positive engagement with an expert. In parallel situations around non-health issues, I've found the ego of the expert seems to be the determinative factor in whether or not the interaction goes well.
This is why the world model approach is so important. It allows you to feed back the prediction accuracy of the model to itself at training time, enabling it to predict (to some degree) its own uncertainty. If you jump through a couple of hoops you can also do this at run time to give it “spidey sense” that something’s not right with current inference.
RMSE is just an extrapolation from the training data. If the data is wrong because the world changed, any model (parametric or not) can be confidently incorrect.
I gave up on Grok. It is going the way of Tesla, SolarCity, gigabattery and Autopilot. Now on GLM5.2 via Open router hosted in SG. Mimo is also good. Their agent is so convenient and Deepseek level cheap. Quality a bit behind GLM5.2. But then Mythos is myth, technically GLM is high up there in quality but on lower end pricing.
My father in law owns a small manufacturing business and is not technical at all. His computer skills stop with some CAD and basic excel. He pays for ChatGPT as does his wife and her kids. The internet and dot com bubble didn't have millions and millions of non technical users paying cash for a product. Almost every coffee shop I go to has people talking about AI and ChatGPT even in areas with no tech populations.
I still think it could crash, but it's got real users and a mind share like nothing I've ever seen.
The internet and dot com bubble didn't have millions and millions of non technical users paying cash for a product
The dot com bubble was basically based on regular people buying computers and internet service, and then using them to buy products they used to buy in stores.
This seems to ignore the fact that millions of non-technical people did pay cash for a product: AOL. And in fact the AOL buyout of Time Warner coincided almost exactly with burst of the dot com boom.
That may be the case but the idea that non-technical people didn't pay cash for internet products before the dot com boom is fanciful. I would say AOL's acquisition of a legacy media company did signal the peak of a mania the way these IPOs are.
These companies are never, ever going to make their money back off of retail customers. It's not even clear if those customers would be profitable at all, let alone enough to justify hundreds of billions in capital expenditures.
The question in my mind is persistence. Everyone goes through the honeymoon phase. I'm absolutely loathing the idea that phones are arriving soon with chatbot junk built deeply into it, enough that the thought is more what if I could maybe just stop using my phone so much. I threw myself at the Llama WhatsApp integration when I first got it, now the idea of having Llama in WhatsApp just feels so dumb.
I was a huge early fan of ChatGPT voice too, but I don't think I've used voice mode anywhere in at least 6 months. The question is what is the right level people are generally going to settle on for the use of these tools in the long term. 80% of my usage isn't much more than a better Google, I could live without it and I could live with cheaper options. I'm not sure the consumer money is going to be there en masse as hoped
Of course it still leaves a huge amount of business cases open, but I suppose the same principle applies. How soon will people tire of talking to robo-voice when they call their bank? etc.
I definitely believe in the broad existence of people like your father in law. What I’m not sure about is how many of them would keep paying if their subscriptions were priced profitably.
reply