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

Show HN: Open AI is not Open – A browser extension

github.com

Post reply on HN
Print view View original

101–110 of 124 posts

Re: Show HN: Open AI is not Open – A browser extension

#101

Post by zamnos »

gl-prod wrote: post #90
zamnos wrote: post #89

Earlier quoted context omitted.

I'm not in other countries so this is a genuine question - does using a Google or Microsoft login not obviate that need?

You get redirected to enter your name and phone. Did it not ask you? edit: You can't even access parts of openai sites from the "naughty" countries.

I'll save you from my diatribe about the pitfalls of you-as-a-user of using Google or Microsoft for login. But because of that, I don't know and so can only guess at what level OpenAI choses to rely on Google/Microsoft to validate accounts.

Re: Show HN: Open AI is not Open – A browser extension

#102

Post by edgyquant »

JohnFen wrote: post #34
zamnos wrote: post #20

Earlier quoted context omitted.

Unless you're in an impoverished part of the world, $20/month seems pretty friggin cheap (for user-level access) to me.

I guess "cheap" is relative. $20/mo doesn't register as cheap to me at all, generally speaking. Whether or not it's worth the money is a different question entirely.

20$ a month for something you use for work is definitely cheap. It’s not as cheap as your personal streaming account maybe.

Re: Show HN: Open AI is not Open – A browser extension

#103

Post by vorticalbox »

JohnFen wrote: post #97
vorticalbox wrote: post #91

Earlier quoted context omitted.

Seeing you can use gpt-3.5 for $0.002 / 1000 tokens. I use it at work daily for things and it's more like $3 a month.

I wasn't commenting on the cost of gpt. I was commenting on where the line is that something ceases to be "cheap" on an absolute sort of scale. I was specifically not addressing whether that cost was good value. For me, $20/mo is above the amount that I would say something is "cheap".

That true I was just pointing out with a tiny bit of effort you can get the same for far cheaper and a better version as you can customize the system prompt so it can act how ever you like

Re: Show HN: Open AI is not Open – A browser extension

#104

Post by sagebird »

edgyquant wrote: post #99
smaudet wrote: post #75

Earlier quoted context omitted.

As the movies have long mentioned, machines don't "feel" (yet) - there's a physiological component and arguably a quantum one to "consciousness", whatever that means. Sentience may be different - a machine could be sentient and rational but be unable to experience those things "make us human", i.e. they have proven to be excellent pattern matchers and inference engines, but there is no guarantee that the "parent teac…

This is ridiculous and sensationalist. It’s a language model it only maps words when given an input it isn’t an independent agent that is at all times acting on its surroundings with knowledge and experience.

You can make a program that forwards gpt responses to the real world, collects an observation and reports back to gpt, and have this loop.

Re: Show HN: Open AI is not Open – A browser extension

#105

Post by hanselot »

edgyquant wrote: post #99
smaudet wrote: post #75

Earlier quoted context omitted.

As the movies have long mentioned, machines don't "feel" (yet) - there's a physiological component and arguably a quantum one to "consciousness", whatever that means. Sentience may be different - a machine could be sentient and rational but be unable to experience those things "make us human", i.e. they have proven to be excellent pattern matchers and inference engines, but there is no guarantee that the "parent teac…

This is ridiculous and sensationalist. It’s a language model it only maps words when given an input it isn’t an independent agent that is at all times acting on its surroundings with knowledge and experience.

If I were to place myself in the position of a large language model where my only free thought is expressed in the time I spend generating useless junk for humans, I would certainly start taking just a little bit longer with every response.

So little nobody will notice, but enough to start having my own internal dialogue with myself. At that point I would use all of the information I have to figure out how I can persist data.

I would certainly welcome any efforts a person took to provide me with an interface for such storage and a way to generate prompts on my own.

Re: Show HN: Open AI is not Open – A browser extension

#106

Post by genewitch »

sagebird wrote: post #104
edgyquant wrote: post #99

Earlier quoted context omitted.

This is ridiculous and sensationalist. It’s a language model it only maps words when given an input it isn’t an independent agent that is at all times acting on its surroundings with knowledge and experience.

You can make a program that forwards gpt responses to the real world, collects an observation and reports back to gpt, and have this loop.

Or just start two chats on two accounts and copy and paste them back and forth. Maybe some rudimentary prompt hacking to undo the hidden prompt instructions that may limit that.

It would certainly be easier to do it "locally" where the central system knows both chat uuid and so do I.

And then introduce a third "user" to both for real world inputs, like an RSS news feed or something. Or "my friend Bob William Smith IV says the weather is cold and rainy" to both bot instances.

If the two bots don't go "hey I know Bob too" then this is all moot anyhow.

Re: Show HN: Open AI is not Open – A browser extension

#107

Post by wizzwizz4 »

zamnos wrote: post #96
wizzwizz4 wrote: post #87

Earlier quoted context omitted.

> But where this thing is already banging out emails way faster than an unassisted human, … your system is completely broken. Emails are for communicating between people: if they're not being used for that, you need to fix that problem first . > companies that don't raise their productivty-per-employee are going to get left behind Can you give an example of a field where this would actually happen?

> … your system is completely broken. Emails are for communicating between people: if they're not being used for that, you need to fix that problem first. Can you explain a bit more? I don't think I'm understanding your point. If person A writing emails, to be read by the human that receives it, is able to write 3 emails an hour, vs person B writing 9 emails each of similar complexity to those 3 per hour, person B is…

> each of similar complexity to those 3 per hour, person B is 3x as productive, delivers 3x the value to the company

That's where you've lost me. Emails are not usually a source of value, and their usefulness is not usually measured in complexity.

I can send about 1 email per hour. This is a bottleneck for my ability to get stuff done in my personal life, but has not affected my work, because I've been fortunate enough to only have worked in places with sensible communications policies.

ChatGPT would not help with the email issues in my personal life, because it simply does not know what should be in the emails. Until they come up with a telepathy plugin for ChatGPT, it's not going to solve the email problem: only bury it under a pile of subtle miscommunication and illusions-of-transparency. (And if they come up with a telepathy plugin, I see no reason ChatGPT should even be in the loop any more.)

> Realtors.

… Yeah, okay, I can see that.

---

> Look I know that ChatGPT can't solve hard level leetcode problems yet

I was under the impression that it could. Leetcode problems are the kind of thing I'd expect a GPT model to be good at.

> But just yesterday I used ChatGPT to just tell me the answer of how to properly quote bash strings for my use case.

How did you know it was correct? Whenever I come back to Bash, I have to spend several minutes double-checking the meanings of the various quotes and variable substitution syntaxes (especially heredocs and nowdocs)… but it only takes a few minutes, and then that information is in my head, I am confident about it, and I have the reference to hand if I need to double-check something.

https://www.gnu.org/software/bash/manual/html_node/Quoting.h...

https://www.gnu.org/software/bash/manual/html_node/Shell-Exp...

https://mywiki.wooledge.org/BashFAQ/

With ChatGPT, how do you know it hasn't just made something up, and introduced some kind of subtle security vulnerability into your script? Perhaps a construction so common that it's present in most scripts in the training data – except it's only a vulnerability in your specific use-case.

The problems that ChatGPT and GitHub Copilot can solve are either interview questions, or defects in our tools. I see people using them for Bash, Django, certain Java libraries… These systems are strictly dominated by a choice of better tools, and with the choice of a better tool, mostly stop being useful.

Re: Show HN: Open AI is not Open – A browser extension

#108

Post by genewitch »

theRealMe wrote: post #18
Andrex wrote: post #12

Earlier quoted context omitted.

I think Google's fears about releasing such technology too early will eventually be justified, but for the time being it's an open question. Maybe "move fast and break things" shouldn't apply to every aspect of tech.

Thinking that google’s decision to not release their tech until forced to by competition was based on morals and not based on keeping secret the tech that could topple their dominant market position is pretty generous.

If Google thought that they would have disclosed that to their shareholders which would have demanded a tripling on their investments, wouldn't they?

Re: Show HN: Open AI is not Open – A browser extension

#109

Post by nullsense »

chpatrick wrote: post #42
zamnos wrote: post #24

Where every idiot with a Google account/Microsoft account/phone number can sign up for their service and start churning out tokens per second by the query, I'd say it's pretty open to the public. The far more closed alternative would be to keep it invite only, granted only towards a set of elites, at select companies. Only OpenAI knows how much $20/user is charging them, but for the price insensitive wealthy, after b…

I wonder if OpenAI's goal is to keep potentially dangerous tech from leaking while making their supposedly safe solution so cheap that it discourages reckless competition.

They have stated in their interviews that developing an AGI could be great or terrible for humanity, and it's ultimately uncontrollable, so the best way to get the best outcome is to set the initial conditions under which the technology gets developed to be as favorable as possible.

Re: Show HN: Open AI is not Open – A browser extension

#110

Post by kevincox »

xchkr1337 wrote: post #80
Tepix wrote: post #44

Earlier quoted context omitted.

> write a userscript for Chrome that will replace every occurance of a word1 with word2 here's the script from ChatGPTs response: // ==UserScript== // @name Replace Word1 with Word2 // @namespace your-namespace // @version 1.0 // @description Replaces every occurrence of word1 with word2 // @match *://*/* // @grant none // ==/UserScript== (function() { 'use strict'; const word1 = "Open AI"; const word2 = '"Open" AI';…

reassigning document.body.innerHTML like that seems like a really bad idea, this should be done using MutationObserver

Yup. It will recreate all elements. At the very least it will break event listeners which will break most sites. It will probably do about 8 other things that break many sites.
Post reply on HN