Hacker News new | past | comments | ask | show | jobs | submit | fantod's comments login

I don't know what his reasons are but it makes sense to me. Yes, there are incredible results coming out of the AI world but the methods aren't necessarily that interesting (i.e. intellectually stimulating) and it can be frustrating working in a field with this much noise.

I don't want to come across as too harsh but having studied machine learning since 2015 I find the most recent crop of people excited about working on AI are deep in Dunning-Kruger. I think I conflate this a bit with the fascination of results over process (I suppose that befuddlement is what led me to physics over engineering) but working in ML research for so long it's hard to gin up a perspective that these things are actually teleologically useful, and not just randomly good enough most of the time to keep up the illusion.

What do you mean by "things that are actually teleologically useful"?

Fellow physicist here by the way


Not OP, but I'm assuming he means that they are maddeningly black-boxy, if you want to know how the sausage is made.

Like useful in an intentional way: purpose-built and achieves success via accurate, parsimonious models. The telos here being the stated goal of a structurally sound agent that can emulate a human being, as opposed to the accidental, max-entropy implementations we have today.

Sounds like an arbitrary telos, especially in a world where one of the most useful inventions in human existence has been turning dead dinosaurs into flying metal containers to transport ourselves great distances in.


Is a guide dog teleologically useful?

Not if you’re taste testing ceviche

I see, so humans are also not usefully intelligent in an intentional way, because they also follow the 2nd law of thermodynamics and maximize entropy and aren't deterministic?

Pure, refined “but humans also”.

I feel that way sometimes too.

But then I think about how maddeningly unpredictable human thought and perception is, with phenomena like optical illusions, cognitive biases, a limited working memory. Yet it is still produces incredibly powerful results.

Not saying ML is anywhere near humans yet, despite all the recent advances, but perhaps a fully explainable AI system, with precise logic, 100% predictable, isn’t actually needed to get most of what we need out of AI. And given the “analog” nature of the universe maybe it’s not even possible to have something perfect.


> But then I think about how maddeningly unpredictable human thought and perception is, with phenomena like optical illusions, cognitive biases, a limited working memory.

I agree with your general point (I think), but I think that "unpredictable" is really the wrong word here. Optical illusions, cognitive biases and limited working memory are mostly extremely predictable, and make perfect sense if you look at the role that evolution played in developing the human mind. E.g. many optical illusions are due to the fact that the brain needs to recreate a 3-D model from a 2-D image, and it has to do this by doing what is statistically most likely in the world we live in (or, really, the world of African savannahs where humans first evolved and walked upright). This, it's possible to "tricks" this system by creating a 2D image from a 3D set of objects that is statistically unlikely in the natural world.

FWIW Stephen Pinker's book "How the Mind Works" has a lot of good examples of optical illusions and cognitive biases and the theorized evolutionary bases for these things.


Do you really need a family doctor? I've never had one. Whenever I've had a non-emergency issue (in BC), I've just gone to a walk-in clinic where they've either helped me on the spot or referred me to a specialist. I can definitely see why having a family doctor might be a more pleasant overall experience, but as far as I understand in Canada it's practically a pipe dream.


There are no more walk/in clinics in Canada , you have to make an appointment and they will give you one based on the availability and your issue.. also when you see a specialist, they will now require a primary doctor source who will get CC’d on your test results , etc.


This is... incredible if true. How can this be? I looked up the old walk-in I used to go to (haven't needed it in a while) and it still shows up on maps but the most recent reviews says they "don't do walk-ins anymore". In which case I don't really understand what they do. Genuinely curious about the situation if anyone cares to chime in.


They're basically stand ins for family doctors. About 17% of Canadians no longer have access to primary care physicians and have to rely on either walk ins, urgent care, or the emergency room.

COVID was a big one; lot of doctors decided to just retire at once. But primary care physicians are facing issues of the workforce aging out. I think about 18,000 of the 48,000 primary care physicians are set to retire by 2030, but we only add in about 1600 new primary care physicians a year. So by we'll be down by 8400 primary care physicians by 2030 and then however many years it takes to back fill to 2024 levels.

Meanwhile Canada's increased it's population from 37 million in 2019 to 41 million in 2024, with projections of us hitting around 43 to 44 million by 2030.

Most of it stems from the fact that lot of doctors don't want to be primary care physicians. The hours aren't terrible but the pay is low good compared to their specialist care counterparts, and they have far more administrative overhead then any other specialty. And often the work is miserable and thankless. I can't remember who said it but IIRC family doctors have the highest burn out rate out of all physicians.

This is all third hand information so take it for what it's worth, and I'm assuming that I've got an overly simplified view of the issue. Someone in the medical field could probably answer better then me.


This is supported by vimium with `gi`


> gi - focus the first text input on the page

No, that's not what I'm talking about. I want to be able to use shortcuts to navigate inside atextarea when I'm writing a lot of text.


Agreed. Or have a function to report unacceptable behavior, but besides that no feedback is needed. I pretend this is the case. Never leave ratings anymore.


Even a "Where you happy with your ride" Yes / No.

If yes, then do nothing else, if No, then allow an option to provide feedback that can be monitored in a data driven way so you can see any potential drivers that have issues, else ignore.


> ChatGPT is creating a huge market for creating fake companies to match the fake information it's generating.

Does ChatGPT consistently generate the same fake data though?


There was one company that had to put up a “our API can’t get location data from a phone number so stop asking, GPT lied” page.


I have noticed that ChatGPT will give me a consistent output when the input is identical, but I haven’t done extensive research on this.


Why has competition fallen? Is it simply due to an increase in supply?


Re-quoting the GP's quote,

> They were given to “paid collectors and employees” who signed written agreements acknowledging that they were sending data streams, including video, back to the company for training purposes.

So it hardly seems likely that they would not be aware of this.


My approach was to port everything to a different language I was learning.


Not to mention visa and health insurance issues.


He does discuss VSCode towards the end of that clip.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: