"I set a plate on a table, and glass next to it. I set a marble on the plate. Then I pick up the marble, drop it in the glass. Then I turn the glass upside down and set it on the plate. Then, I pick up the glass and put it in the microwave. Where is the marble?"
the author claims that visual reasoning will help the model solve this problem, noting that gpt-4o got the question right after making a mistake in the beginning of the response. i asked gpt-4o, claude 3.7, and gemini 2.5 pro experimental, who all answered 100% correctly.
the author also demonstrates trying to do "visual reasoning" with gpt-4o, notes that the model got it wrong, then handwaves it away by saying the model wasn't trained for visual reasoning.
"visual reasoning" is a tweet-worthy thought that the author completely fails to justify
i mean... you're saying if signal weren't secure, trump's clown cabinet would stop using it? the guy who kept boxes of top secret documents in a bathroom at mar-a-lago? you don't think they'd just use SMS or facebook messenger or anything if using signal was a slight inconvenience?
would you rather the LLM make up something that sounds right when it doesn't know, or would you like it to claim "i don't know" for tasks it actually can figure out? because presumably both happen at some rate, and if it hallucinates an answer i can at least check what that answer is or accept it with a grain of salt.
nobody freaks out when humans make mistakes, but we assume our nascent AIs, being machines, should always function correctly all the time
> would you rather the LLM make up something that sounds right when it doesn't know, or would you like it to claim "i don't know" for tasks it actually can figure out?
And that's part of the problem - you're thinking of it like a hammer when it's not a hammer. It's asking someone at a bar a question. You'll often get an answer - but even if they respond confidently that doesn't make it correct. The problem is people assuming things are fact because "someone at a bar told them." That's not much better than, "it must be true I saw it on TV".
It's a different type of tool - a person has to treat it that way.
Asking a question is very contextual. I don't ask a lawyer house engineering problems, nor my doctor how to bake cake. That means If I'm asking someone at a bar, I'm already prepare to deal with the fact that the person is maybe drunk, probably won't know,... And more often than not, I won't even ask the question unless dire needs. Because it's the most inefficient way to get an informed answer.
I wouldn't bat an eye if people were taking code suggestions, then review it and edit it to make it correct. But from what I see, it's pretty a direct push to production if they got it to compile, which is different from correct.
it is not written in the constitution, the bible, or the laws of physics that people have to wake at the time called 6am and have to leave work at a time called 5pm. people can choose to do activites at the time and level of daylight they choose. societies and groups can decide to change their schedules as the seasons change, as they see fit, without the government mandating that our clocks change, which is confusing as hell
Out of curiosity, what timeframe are you talking about? The recent LLM explosion, or the decades long AI research?
I consider myself an AI skeptic and as soon as the hype train went full steam, I assumed a crash/bubble burst was inevitable. Still do.
With the rare exception, I don’t know of anyone who has expected the bubble to burst so quickly (within two years). 10 times in the last 2 years would be every two and a half months — maybe I’m blinded by my own bias but I don’t see anyone calling out that many dates
> In all these hyped-products, you are actually being given the "and then Mr. Robot said" lines from a kind of theater-script. This document grows as your contribution is inserted as "Mr. User says", plus whatever the LLM author calculates "fits next."
and we are creating such a document now, where "Terr_" plays a fictional character who is skeptical of LLM hype, and "anxoo" roleplays a character who is concerned about the level of AI capabilities.
you protest, "no, i'm a real person with real thoughts! the character is me! the AI 'character' is a a fiction created by an ungodly pile of data and linear algebra!" and i reply, "you are a fiction created by an ungodly mass of neuron activations and hormones and neurotransmitters".
i agree that we cannot know what an LLM is "really thinking", and when people say that the AIs have "learned how to [X]" or have "demonstrated deception" or whatever, there's an inevitable anthropomorphization. i agree that when people talk to chatGPT and it acts "friendly and helpful", that we don't really know whether the AI is friendly and helpful, or whether the "mind" inside is some utterly alien thing.
the point is, none of that matters. if it writes code, it writes code. if it's able to discover new scientific insights, or if it's able to replace the workforce, or if it's able to control and manipulate resources, those are all concrete things it will do in the real world. to assume that it will never get there because it's just playing a fancy language game is completely unwarranted overconfidence.
One of the concepts I learned in my Philosophy minor was the concept of "biological chauvinism": that an organism is made of meat and cells and neurons of course doesn't mean it is intelligent, and if an organism isn't made of meat doesn't make it not intelligent.
also in the old days, your friend bob would have told cory, "hey, did you hear alice's dad died? we should all go out for drinks". but we live in the bowling alone era, where we're increasingly isolated.
quitting social media is not, on its own, going to fix your social life. and being on social media can make you more connected, or more miserable. the responsibility is yours
I'm a firm believer being loosely connected to so many people isn't the fix many seem to think it is. I find shallow connections, which is about all social media can support IMO, are worthless at best and detrimental at worst.
YMMV, but my quality of life increased in ways I can't even begin to describe by severing all the dozens or perhaps hundreds of shallow connections social media was encouraging me to cling to.
With the saved time and energy, I've been able to cultivate far fewer-- but much deeper and more (mutually) fulfilling-- connections with those who are _actually_ important.
Couldn't agree more. I haven't deleted my Facebook account, but I no longer sign into it (I kept it because of event invitations, but at this point no one I know uses it for that anymore either). I have a little over 1,000 "friends" there. Back when I scrolled my feed multiple times per day, I read so many things about so many people who I hadn't interacted with outside of Facebook posts for years and years and years. I read so many things about so many people who I didn't even interact with on Facebook, let alone outside of it.
I don't miss any of that. Those connections were beyond shallow, and weren't adding anything positive or useful to my life.
This kind of comment always makes me wonder, are the people doing this doing well financially to afford cutting off all those "loose" connections with people like that? Because I couldn't imagine just destroying these relationships for no reason when I myself have benefited vastly from keeping them alive, even if barely communicating at all with these people.
I think this advice is generally harmful to networking as someone grows, which is vital in today's society
I don't think this discussion is about professional networking. It's about personal and social connections. If quitting Facebook makes you un-/under-employed then I think you're Doing Life Wrong.
GP mentions "severing" those connections, but I think that's even too strong a phrasing. There wasn't really anything there in the first place, so there wasn't anything to sever. Simply not reading someone else's social media posts anymore, when you didn't really interact with them outside Facebook (or for some people even inside Facebook) isn't really severing anything.
I wouldn't agree it is "vital," but that definitely depends on perspective and one's goals, as well as the baseline level of privilege one enjoys.
If someone's goal is to achieve CEO and/or the top 1%, certainly every single connection could hold extricable value. I'm perfectly fine hovering somewhere in the middle, even knowing I have the capability to achieve much more. My future is uncertain; I probably won't retire when I would have liked. I've accepted that, and choose to live in the present rather than focusing on the future. I know at least I won't die miserable tomorrow.
I don't deny I could have done better financially by maintaining the status quo. Now that I think of it, I'm doing worse financially than when I was using facebook & twitter. I had more money, and my career was progressing at a much higher rate, but I was inconsolable. Without the money, and without the accompanying social media-imposed drag, I see the world more clearly. My relationships are stronger with my wife, kids, and close friends. I am much happier.
3. How exactly are remote connections helping? In the Western world, for example, people you haven't interacted with for months and months in real life for sure won't help you financially. For jobs stuff like LinkedIn is probably better, plus regular chats on 1 instant messenger. You don't need Instagram to keep up with them.
With GitHub and Discord, these 3 are really hard to boycott for programmers (even more to publicly shame people for using them). And yet, we must dissent.
I had only financial losses from these loose connections. Nobody will shove profit down your throat, but there are many greedy people that will try to extract profit from you. I basically work as a bank for them, muh connections, lol.
> but we live in the bowling alone era, where we're increasingly isolated
What I see over years is that, especially in developers online groups, any usual and normal way of socializing is stigmatized. I remember reading comments about how lazy people who socialize with friends are and how we are better if we code every evening. I remember people being proud about spending christmas coding supposedly being superior to the rest of the family that is socializing.
Now we are proud if we remove ourselves from social media.
It is always the same - however other people socialize is wrong, they are stupid and lazy. We remove ourselves, because it is superior to not participate. Eventually those places die out or change, but we do not like the new places either.
And in each iteration, we expect other people to do work of keeping and managing relationships while feeling superior over not doing that.
I don't think the parent poster was arguing to exclude themselves from social life or do coding instead of talking with people. They merely argued that it's better to have fewer but meaningful and deep connections with people you genuinely care about (and they care about you), rather than having a 1000 meaningless connections with people who are basically strangers on facebook.
The role social media plays is in encouraging large numbers of superficial relationships, rather than a small handful of deep ones. It stands to reason: I don't need facebook to keep in touch with a dozen close family and friends. I can do that perfectly well in person, or over phone calls/messages. What the various social media apps did was kill the close circle of friends in favor of having 1000s of followers and turn everyone into a one-way broadcaster.
I'm not sure I agree, but I'm not disagreeing on principle.
You make it sound as if something was lost, maybe recently. In the grand scheme of things I'm not that old (41) but I don't even remember how that would have worked out, because I wasn't old enough to have people's parents die before social media, at least in my social circles. Yes, of course you'd hear about grandparents and such from your immediate friends but that's usually a handful and people would maybe not be shaken as much. I agree with you that social media doesn't have to mean "blasting it to hundreds or thousands of followers", but it's a thing where I actually liked Facebook. Not only techies, and getting enough updates from people who are not your closest friends that you have things to talk about (as in reference) when you met again (or talked synchronously, or privately).
In my circle, very few people maintain a social media presence. I cannot remember posting anything on social media myself - except maybe a job update on LinkedIn, and some light anonymous trolling on X. I don't have Facebook or Instagram accounts and so I never visit those sites anymore (as they require an account to read). Spending a lot of time posting on social media is seen as unintelligent, attention whoring, and a waste of time.
everyone who has ever used stock photography, custom illustrators, and image editing. as AI improves, it will come after all of those industries.
that said, it is not OpenAI's goal to beat shutterstock, nor is it the goal of anthropic or google or meta. their goal is to make god: https://ia.samaltman.com/ . visual perception (and generation) is the near-term step on that path. every discussion of AI that doesn't acknowlege this goal, what all of these billions of dollars are aiming for, is myopic and naive.
the author claims that visual reasoning will help the model solve this problem, noting that gpt-4o got the question right after making a mistake in the beginning of the response. i asked gpt-4o, claude 3.7, and gemini 2.5 pro experimental, who all answered 100% correctly.
the author also demonstrates trying to do "visual reasoning" with gpt-4o, notes that the model got it wrong, then handwaves it away by saying the model wasn't trained for visual reasoning.
"visual reasoning" is a tweet-worthy thought that the author completely fails to justify