Comment by dang
2 days ago
As an aside, the variety of examples in that other thread is impressive. Here are some that I noticed:
Fixing my furnace: https://news.ycombinator.com/item?id=48420635
It just goes on and on. I was a little nervous when I saw that post originally, but it's amazing what happens when a title is somehow just right.
Oh dang, we all know –based on observation of your throughput and availability– that you're AI, you just can't "be nervous".
Jokes aside, thanks for your selection. I had read some, but missed others until your comment.
If it matters, I think there's some people that hasn't decided yet what tribe (pro-AI/anti-AI) they belong to. There's probably dozens of us!
The argument is dialectical. In other words both sides are right.
There is a really important question that is been lost between the anti and pro AI camps which is really answering what AI is good at and what AI is bad at, and what is the root cause of the weaknesses. Is it intrinsic to the models that LLM's use, or the way thy have been trained. In this knowledge is where the gold mine is for the next start up.
For example AI is very good at answering well defined questions, but suffers from premature closure. It will not know if it has all the information to answer the question. So whilst AI will score better than a doctor or a lawyer on a domain question it will not necessarily gather all the evidence needed to answer the question properly. Knowing this whilst using a LLM is a super power.
There is also a large gap Usability issue in that often the LLM does not really know the humans context, in other words a context collapse. It does not know if you specilised in your domain or just asking for fun.
We should be exploring and debating these gaps.
> There is a really important question that is been lost between the anti and pro AI camps which is really answering what AI is good at and what AI is bad at, ...
I think some of this pro/anti AI discussion is really a proxy for something else though. There is another unstated, societal disagreement about personal responsibility and the behaviors of the AI tool users.
It isn't entirely new, as people have always been blaming abstractions which lack agency (society, bureaucracy, "the machine", "the process", "the game", etc.) to conduct themselves in an ethically or logically flawed manner and then shirk responsibility. But, I think it is accelerating dramatically with the way AI agent usage is being adopted. I don't see how this can possibly continue to be normalized without leading to a catastrophic outcome.
Right now, AI tool usage is being applied in ways that would have been considered misconduct, negligence, or fraud if the actors were doing the same thing via traditional outsourcing. People who have a contractual and ethical responsibility to apply analysis, judgement, and oversight are rapidly turning into dumb pipes who just relay content. They essentially phone it in, trying to take credit for apparently moving KPIs "up and to the right". They pretend to themselves and others that they are reviewers and still in control, but when things go wrong they expect forgiveness and for others to lean in and clean up their rapidly spreading mess.
In software circles, many of us are appalled by this slop that is poisoning the well for our existing organizations. It seems like a massive, toxic externality. To us, people aiming their AI agents into our existing collaborations are pretty much flipping a switch to "roll coal", or worse.
7 replies →
There are also madman who get tools and capabilities via AI sane society would never hand anyone. Thos is a liberation for the untamed fringe.
I haven't had the patience or mentality to really absorb "ok but how is this useful" until that thread and your highlighted references. Thank you for the curated highlights, however brief it may be, because it's very hard to find such diamonds without dedicating far too much time wading through the abstract gatekept comments on the topic, in most cases. Real world examples give me much hope!
Lemme give another. I was formerly quite anti-AI but bought a cheap Claude plan just to play around with it a bit. First thing I built with it was this - https://github.com/tylereaves/onscreen-piano, in about an hour and maybe 10 prompt cycles. It replaced, for my specific use case, the 10% of the functionality of an increasingly-unreliable commercial app. That's including building the website, setting up actions for mac and windows builds...
My next project was a 2d game with random terrain, physics, sound, music, multiple levels, a day/night cycle with transitions high score tracking... (not uploaded anywhere, but it works, and I refined it a good bit.). That was more like 8 hours and maybe a 100 prompts.
Here are a few screenshots:
https://imgur.com/a/vhUXBu3
One thing that I have found to make a pretty big difference is using both the latest models and higher thinking levels. Opus 4.8 with thinking on Extra or even Max is genuinely mind blowing. The thing I hadn't really appreciated, having a sort of naive impression formed mainly from using free early versions of stuff like ChatGPT and Stable Diffusion was sort of that "Type a big ass prompt and it craps out a result" experience. But Claude is really great at refining from feedback, and it's way more flexible and responsive than I would have ever expected. I can do something like take a screenshot of a small portion of the running app or website or whatever and just say "This button needs to be bigger" or "make this red" or something like that, or even sometimes just "fix this", and Claude both correctly identifies what I'm talking about, and actually does the thing.
here I've found it really, incredibly game changing is my health. I have a pretty, to put it mildly, complex medical profile at this point. I haven't worked in over a year and pretty much every sign is pointing towards permanent disability at this point. Tons of symptoms, long med list, and I live in a smaller town with not great access to care. I'm also autistic and have not the greatest verbal communication, especially under stress or time pressure. I dumped all my info at it, in bits and bobs over several days (Side note... it's memory is pretty limited, but it will quite happily right out everything it knows from a session into a markdown file it can later re-read. I've found it very good for things like screening for drug interactions, or talking through and logging symptoms (and it can log those into human readable markdown files too). Biggest win (other than having unlimited time and interactions) is that it thinks across specilaties, versus the "real world" where the gastro only wants to deal with gastro stuff, neurology only wants to do neuro.
I certainly don't (and wouldn't) use it as a replacement for a doctor, but as an adjunct it's phenomenal. For instance, it flagged a possible drug interaction with a symptom I was having, and then offered to draft a portal message to my GP about it. I have poor executive function so lowering the friction from "type up a message and send it" to "copy and paste" is actually a pretty big deal. Turns something (I probably won't do) later into something I will do now.
It wouldn't surprise me if my very direct, literal, autistic communication style is particularly well suited to interacting with AI. I actually find talking to it rather refreshing as, while of course it's not perfect, it tends to actually respond to what I say rather than the all the assumed subtext NTs tend to expect/react to.
I am very optimistic about people using LLMs to explore their medical issues. Human bodies are complex and doctors are time- and intelligence- limited. An LLM can be a really useful partner.
2 replies →
Your last paragraph hit home with me. I also find it refreshing to be clear and direct, have the intent be directly understood, and have a pleasant, good faith, high signal conversation as a result. I hardly ever experience that with humans (I recognize that that’s not ‘everyone else’s fault’ of course).
I always thought of myself as merely ADD, but your comment is so incredibly relatable across all dimensions that I find myself wondering how far along the autism spectrum I am. Although, in my personal opinion, neither term is especially helpful: it's not a sensory/executive processing disorder if it's clearly benefiting both yourself and the people around you. It might make being sociable difficult, but that's the tradeoff for being willing to engage your mind in ways that others filter out as "too much information."
3 replies →
On the medical pieces, a very nice agentic medical AI worth looking at is https://biomni.phylo.bio
[flagged]