Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I'm a seeing-eye dog for a computer

claytonwramsey.com

Post reply on HN
Print view View original

61–70 of 105 posts

Re: I'm a seeing-eye dog for a computer

#61

Post by hypfer »

walrus01 wrote: post #59
hypfer wrote: post #40

Earlier quoted context omitted.

> And at least 55-75% of the people know each other face to face through industry vertical specific events, trade shows, conferences and similar. Tbh, my experience with _that_ has been worse than with a random sample of internet background radiation, as people seem to have no idea how to filter correctly and who to vouch for and why. _Especially_ with people they know via work, conferences, etc. Spaces that do not f…

Within this context it's specifically people in the ISP and telecom business, most of whom have been in the business 15, 20 or 25 years now. Some of us started out as junior NOC techs, junior sysadmin/Linux/BSD guys in the late 1990s at different dotcom 1.0 startups. At a certain scale you know who the people are that have super-admin privileges at very specific ASNs and they are well known in the industry. Basically…

Yes, I got that, but what I actually mean is that these criteria do not necessarily lead to good conversations.

I know the kind of space you are referring to. I'm in those too. But my point is that they regularly disappoint me.

Which is usually worse disappointment than with randoms on the internet, because the people should know better but don't.

With these tight knit spaces you get all the social dysfunctions of people liking each other not for merit but because they've been around each other for long, and that's - for me personally - worse than the public alternative, as you suddenly do not argue with logic, but with the tribe and social cohesion.

Re: I'm a seeing-eye dog for a computer

#62

Post by Fr0styMatt88 »

I've found that LLMs are specifically bad at a certain kind of debugging, though I can't quite put my finger on what that is.

"Spot the bug in this code" when the code can be looked at and pattern-matched against bugginess is something they seem really good at.

Some parts of debugging, like "Here is this logfile, what do you think is going on?" are also surprisingly good.

It's that thing kind of in the middle -- I know it when I see it honestly is the best way I can put it into words. An example from recently, I'm receiving some bad data on a network message parser. Immediately I don't know whether it's a my-side or their-side thing, but I know if I try and just vaguely describe the behaviour to the LLM it will start churning tokens.

My current approach to problems like this is -- I need to tell the LLM what it needs to do to give itself the data it needs to solve the problem. My first reaction now isn't "It's not working, there's a bug, it's not doing X". It's "Okay, this isn't quite working properly; I need you to add some debug logging around X, Y and Z so we can figure this out". That tends to avoid spirals and get me out of the situation much more quickly.

The seeing eye dog analogy is pretty apt actually. I would love to see some transcripts from the author if they are able.

Edit to add: I think the 'thing' I'm alluding to might be -- if I have trouble expressing the buggy behaviour clearly in words, then I know it's probably going to be a fair few back-and-forths with the LLM to get something; the harder I find it to concisely describe, the more risk that it'll fall into a pit. Doubly so if I offer up a hypothesis which turns out to be wrong.

Re: I'm a seeing-eye dog for a computer

#63

Post by arjie »

skrebbel wrote: post #43
arjie wrote: post #34

Earlier quoted context omitted.

Haha it’s just a foolish move made by those from the ‘90s. We all internalized a different lesson. The younger generations have learned to brush it off and stop responding. Even if the person is human, when you fail to move toward Aumann Agreement, there is no point in persisting. Us older ones are still fighting the last war. If only we explain it properly, maybe the bots will understand!

I googled Aumann Agreement and it’s all about agreeing about whether things are true/false/likely/unlikely. I don’t understand your point at all in this context. What’s “Aumann Agreement” in the context of a discussion about values?

Haha, it’s figurative and requires squinting. “rational actors with common axioms will reach shared conclusions through repeated exchange of information”.

So if you’re not reaching a shared conclusion then you’re not exchanging information, you don’t have shared axioms, or one of you isn’t rational.

If it didn’t land with you, it was just a poor choice of phrase and while I care to attempt to explain it, I don’t care to attempt to defend it.

Values are not different from other positions that people might hold. I believe values and principles are just lossy compressions of positions about specifics because we cannot express them all. When you poke at the edges you find they are fractal. So “what’s your estimate for random variable X?” And “are open borders good?” Are not really different questions so much as questions about different levels of specificity.

My incomplete self-note about this here https://wiki.roshangeorge.dev/w/Principles_Are_Compressed_Im...

Re: I'm a seeing-eye dog for a computer

#64

Post by walrus01 »

ErroneousBosh wrote: post #60
walrus01 wrote: post #35

Earlier quoted context omitted.

I have found that the only way to deal with this these days is basically to only use my time to discuss things with people in private, invitation only forums. HN is a rare exception for "I am pretty sure you all are not bots (yet)" and yet anonymous. Everything else that I care about is in private Signal groups or an equivalent of same with anywhere from 15 to 200 people, and adding any new member means they need to…

> discuss things with people in private, invitation only forums One of the great things about not having ads on rangerovers.pub is that Google is just not interested in indexing it, because it doesn't make them any money. It shows up here and there in search results, sporadically, especially if you know a username or phrase specific to the site, but other than that if you want to know about it you need to be told abo…

I have seen a bit of a resurgence in your classic 2005-era type interface php bulletin board in a browser, as people relegate their facebook/instagram/whatever accounts to only keeping up with family stuff posted by boomers who won't use anything else, or things like niche neighborhood based buy & sell groups (or community notice groups in like a town of 12,000).

I guess there's at least 3 or more active and currently developed software packages with open source licenses, the flaskbb that site uses, simplemachinesforum, the classic phpbb, probably some others I've forgotten about... The LAMP stack dependencies for all seem very mundane and very lightweight.

Re: I'm a seeing-eye dog for a computer

#65

Post by neuroticnews25 »

blauditore wrote: post #54
arjie wrote: post #34

Earlier quoted context omitted.

Haha it’s just a foolish move made by those from the ‘90s. We all internalized a different lesson. The younger generations have learned to brush it off and stop responding. Even if the person is human, when you fail to move toward Aumann Agreement, there is no point in persisting. Us older ones are still fighting the last war. If only we explain it properly, maybe the bots will understand!

Arguing on the internet is not about convincing the other side. It's about winning the majority opinion and getting more votes. Almost like politics.

Obvioulsy not true for all the arguments, only a subset of those not worth having.

Re: I'm a seeing-eye dog for a computer

#66

Post by rcxdude »

onion2k wrote: post #23

One of the first things I tell the junior/mid-level developers I mentor is "You can't debug something just by reading the code." We all have a mental model of how our code works, and it's usually a bit wrong. Bugs are the real world manifestations of those mistakes. When you read the code it's all filtered through your model, and that makes you blind to seeing why something unexpected happened. In order to debug some…

Hmmm, I don't think that's necessarily true. Often times once I have witnessed a bug, I have found it just by reading through the code with the behaviour of the bug in mind.

For LLMs, this is likely to be disproportionately effective as well: especially because they don't really build up a persistent view of the codebase, they're generally re-reading it each session, and they tend to be surprisingly good at predicting the behaviour of code.

(That said, knowing where and how to gather more evidence to make things clearer is a pretty core skill in troubleshooting, so it's generally good advice anyhow)

Re: I'm a seeing-eye dog for a computer

#67

Post by sznio »

dwedge wrote: post #27

> I used to argue with people on the internet, after about six replies, you realize that you’re speaking to someone incapable of thought He realises the current woe of things but doesn't realise he's been arguing with bot farms and teams of people hired just for this reason - to sow doom, arguments and engagement. Around 10 years ago I noticed this happening on trending topics of Twitter - it wasn't that the opponent…

>it wasn't that the opponents were stupid because they disagreed, it was that they were simultaneously intelligent and stupid in the way they spoke in a way that I realised I'd never seen in genuine people, making me realise it was probably different people or bots under one account.

I believed they were all bots until I actually interacted with such a person in real life. It's terrifying.

Re: I'm a seeing-eye dog for a computer

#68

Post by sznio »

novok wrote: post #45
walrus01 wrote: post #35

Earlier quoted context omitted.

I have found that the only way to deal with this these days is basically to only use my time to discuss things with people in private, invitation only forums. HN is a rare exception for "I am pretty sure you all are not bots (yet)" and yet anonymous. Everything else that I care about is in private Signal groups or an equivalent of same with anywhere from 15 to 200 people, and adding any new member means they need to…

That feels like the intellectual equivalent of limiting your career opportunities within a small tiny villages vs. a large city. There is so much more in the city.

I have moved back to my small hometown, from a large city.

My mental health improved significantly. I actually feel valuable at work. I feel pride in contributing to my local community. And the paycheck isn't actually worse.

Re: I'm a seeing-eye dog for a computer

#69

Post by skrebbel »

arjie wrote: post #63
skrebbel wrote: post #43

Earlier quoted context omitted.

I googled Aumann Agreement and it’s all about agreeing about whether things are true/false/likely/unlikely. I don’t understand your point at all in this context. What’s “Aumann Agreement” in the context of a discussion about values?

Haha, it’s figurative and requires squinting. “rational actors with common axioms will reach shared conclusions through repeated exchange of information”. So if you’re not reaching a shared conclusion then you’re not exchanging information, you don’t have shared axioms, or one of you isn’t rational. If it didn’t land with you, it was just a poor choice of phrase and while I care to attempt to explain it, I don’t care…

Appreciate the writeup, thanks.

Re: I'm a seeing-eye dog for a computer

#70

Post by neuroticnews25 »

arjie wrote: post #63
skrebbel wrote: post #43

Earlier quoted context omitted.

I googled Aumann Agreement and it’s all about agreeing about whether things are true/false/likely/unlikely. I don’t understand your point at all in this context. What’s “Aumann Agreement” in the context of a discussion about values?

Haha, it’s figurative and requires squinting. “rational actors with common axioms will reach shared conclusions through repeated exchange of information”. So if you’re not reaching a shared conclusion then you’re not exchanging information, you don’t have shared axioms, or one of you isn’t rational. If it didn’t land with you, it was just a poor choice of phrase and while I care to attempt to explain it, I don’t care…

Have you heard about the free energy principle and free energy model of emotions? In this model, all brain does is minimize the prediction error, so all values and beliefs are bayesian priors by necessity.
Post reply on HN