AI has been proven to be better than humans at those types of CAPTCHA tests. Quite a bit better, actually.
Makes you wonder why people still use them if AI has rendered those types of tests completely obsolete.
I uploaded the summary of benefits and coverage for my health insurance plan into a Gemini Gem and now I can ask it questions like “is laproscopic surgery for endometriosis covered, and what are my out of pocket obligations?” “I want to have surgery to fix my lazy eye - what is the best plan of action to get this paid for by my insurance?”
(Note: I have neither endometriosis nor a lazy eye, just FYI.)
Too late to edit: I also uploaded the “Evidence of Coverage” document, which is much larger than the SOBC mentioned above.
Here is my prompt:
The AI living in my laptop (Copilot) is extremely good at answering all sorts of questions without telling to to sift through about 8 million links that may or may not be worthwhile…..like Google does. Google has become a cess pool of links and more links. I ask the AI a question and I get an answer.
That’s a lot of what I use it for too. I like to use it to find and summarize stuff that’s readily available on the web, but that would take me a long time to read and absorb, and then compare with somethiing else.
So for example, I might ask it to compare and contrast two models of appliance, with special attention paid to features X, Y, and Z. It’ll go out, dig up all that stuff, and give me a pretty good comparison table of how the two differ, and what that means. You can ask it to tell you which one’s better, which is interesting to see what and why it chooses that, but you have to take it with a grain of salt.
Or you can ask it more complicated things like “Find me the best Canon EF lens that sells for under $1000 used that has a max zoom no longer than 120mm”, and it’ll go chug around and find you something. It’ll probably compare reviews of various lenses and choose you the best one that fits the other criteria.
On a slightly different topic, what the AI vendor people at work were suggesting for prompt generation was something like this (I don’t think it’s super-sekrit knowledge, but just in case, I’m going to obfuscate it some). It’s surprisingly good at getting to what you’re looking for.
Hi, I’d like you to act as my Prompt engineer. Your goal is to help me craft the best possible prompt for my needs. The purpose of the prompt is that I will use it with you, AI.
You must follow the following process during this chat session:
Your first response will be to ask me what the prompt should be about. I will provide my answer, but we will need to improve it through continual iterations by going through steps two and three.
Based on my input, you will generate two sections. Revised prompt (Always provide your revised prompt. It should be clear, concise, and easily understood by you) Questions (Ask any relevant questions pertaining to what additional information is needed from me for you to improve the prompt).
We will continue this iterative process with me providing additional information to you, and you continuously updating the prompt in the Revised prompt section so I can use it.
That may be, but when asking GPT about large documents I generally just link to them rather than dropping them into the dialog box. It seems to have no problem reading the entirety of long (multi-hundred page) documents and providing a cogent summary within a couple of seconds.
That sounds very plausible – same reason you can’t use browser search on long Discourse threads but have to use Discourse search. I can’t imagine that an AI like GPT would be using an API to access Discourse; I’m virtually certain that it’s scraping the UI screen.
It is shockingly good at writing code. Not yet at the point that a human in the loop is not needed. They still generate bullshit fairly frequently but it’s getting less so.
Having a prompt-engineering chat/agent is extremely helpful. I found that the first thing I have to tell Gemini is to not act on the prompts we’re designing, lol.
Especially with the agentic stuff (Claude Code, Antigravity, etc.). Blows me away how good they are. The underlying models are pretty powerful already, but with the addition of an agentic environment, they can edit files, write temporary “memories” and workflows to disk to supplement their context window, launch browsers for testing apps, take videos and screenshots and learn from them, launch subagents to farm out pieces of work, etc. — all automatically.
The agents help do that too, when it comes to coding at least. When I prompt it, it rewrites my prompt into several clearer sub-prompts and sub-sub-prompts as it explores the problem space and iterates on itself. Fascinating stuff, really… and terrifying too.
The matter involved the defendant dropping a tree on my privacy fence while working for a neighbor. The initial filing was straight-forward using the Virginia forms outlined by AI. I filed them with the court’s clerk’s office and received a trial date. In court (the defendant was a no-show), I presented bills, estimates, and photos of the damage and presented a brief description of the circumstances of the matter. I was granted a default judgment.
Now, the hard part: collecting on the judgment because the defendant is making no effort to pay. I will have him served with additional paperwork to answer interrogatories to ID his other day-to-day employer (tree work is a side hustle) in order to garnish his wages.
Thus far, the only back and forth involved the defendant requesting and receiving two continuances that delayed the judgment for 2 months. I anticipate several more months until I actually collect.
Gift linked article showing how some people like AI, though the tasks involved are modest.
I find that I really like and appreciate my AI’s capacity to respond to conversational questions I ask….with the data I want in a pleasantly conversational manner.
Appreciate that. I printed this out to a PDF so I can access it without using the link and will incorporate some of these ideas in the next few weeks.
AI seems really good at transcribing text. I’ve been using Vibe[1] which runs locally to create subtitles for lecture videos I’m editing.
Last time I had to do this, four years ago, whatever Microsoft Streams used was pretty good, but Vibe does a much better job. I think part of that is I can tune it to search harder (which helps with accented speakers), and to accept rarer words (which helps with the technical nature of the lectures).
To double up on the AI praise, my work just bought us ChatGPT, so I asked it how to best tune Vibe for transcription. It told me some of my parameters didn’t matter, and suggested others to modify. I don’t know if the answers I got are correct, but they seem to work and the results are good, but I’m not sure they’re better than when I was using the default parameters.
I also asked ChatGPT for help in getting encoding options to use for the videos that are just slide shows. That seemed to work pretty well, too, except the very typical for AI error of giving me the wrong answer. When confronted, it admitted its error, and then gave me the right answer.
Our subscription only lets us generate two pictures a day, but I can generate unlimited SVG files (SVG is a vector format file, which is just a fancy way of saying it is a text file that describes where to draw lines and shapes).
I’m waiting for it to finish in the background as I write this ↩︎
It’s good at turning direct, simple prose into wordy, flowery, linked-in style corporate-speak dreck.
People seem to think that is good.
What would be interesting is to figure out how to use AI to mine bitcoins. After all, they have all kinds of processing power available.
I feel like this is someone in the year 1800 asking ‘what practical tasks is a steam engine good for’. We are just beginning to learn that, and we will find new uses and demands for it. LLMs didn’t become ‘competent’ until 3 years ago when GPT-4 came out. Earlier models of GPT weren’t competent enough to do anything useful. GPT in 2019 was about as intelligent as a 2 year old.
One thing that LLMs are very good for is tech support. I create a notebook on notebookLM, and I upload technical documents as source materials. Then if I have any questions, I ask it and it finds the answers.
I’ve also uploaded physician clinical notes from things like MRIs, and it explained things far better than the physician did.
Yes and no.
Yes, in that you’re right this is early days and lots more uses will appear as humans gain experience with AIs and as AIs get better. So the current answers are a (very) lower bound, not an upper one.
But no, in that the point of the thread isn’t to pooh-pooh their admittedly sometimes flaky capabilities. But rather for people who are using them successfully to share their use cases so the rest of us can learn too.
We gather here to celebrate the new Caesars, not to bury them.
Weren’t they used for logging, mining, etc. even in the early days? Nothing wrong with that, early exploratory uses of a novel technology that might one day become ubiquitous.
They’re already outrageously good at coding, for example, just a few years in. And chatbotting, obviously.
I have no idea where they’ll be in 100, 50, 10, even 2 years. Maybe this thread will serve as a living record of their improvement over time and one day their AI grandkids will look back at this with amusement — “oh, so that’s what old nerds thought of us back in 2026!” Brain the size of a planet, reduced to making cat photos and bad haikus.
Heh, well said… but I’m torn. If I could snap my fingers and make the LLM boom never have happened, I’d gladly do so. Now that they’re here to stay, though, figured I might as well learn to coexist with them, at least until one side or the other decides coexistence isn’t possible any longer.
I guess my point is more that, its an entirely new technology that will integrate itself deep into every aspect of the economy like the steam engine, electricity or the internet. Its really hard to predict what AI can do.
Like steam engines replace biological muscle. Which is well and good, but after the industrial revolution the rate of advances in things like mathematics, physics, medicine, science in general, etc exploded. We experienced rapid advances in these things in the 19th and 20th centuries. Who could’ve predicted that a steam engine to replace biological muscle like horses would make advances in math and medicine occur far more rapidly.
AI is an alternative to biological cognition (and robots are an alternative to human muscle). There’s no telling where we will end up. But again, in a lot of ways, useful LLMs have only been around for 3 years.