News for hackers. Stuff that matters.
Re:The New York Times is suing OpenAI and Microsoft for copy... by MeImCounting on Wednesday December 27, 2023 @01:54PM (#38785210)
"Prompt:

Tell me briefly about the TCP/IP protocol and provide links to relevant sources

ChatGPT:

TCP/IP Protocol Overview Definition & Purpose: TCP/IP (Transmission Control Protocol/Internet Protocol) is the foundational communication protocol suite for the internet and other networks. It dictates how data is transmitted over the network by breaking it into packets, addressing them, transmitting them, and ensuring they are received correctly.

Key Components:

TCP (Transmission Control Protocol): Ensures reliable delivery of data by establishing a connection and managing the transmission process, including packet sequencing and error handling. IP (Internet Protocol): Handles addressing and routing of packets, ensuring they reach the correct destination across diverse networks. Functionality:

Data Handling: Data is divided into packets, which are then sent and reassembled at the destination. Error Correction & Sequencing: TCP handles errors and maintains the correct sequence of data packets, ensuring integrity in communication. Importance: TCP/IP is critical for setting up the internet and intranet networks, allowing different types of computers to communicate effectively regardless of their architecture, operating system, or data structure.

Further Reading & Sources:

RFC Documents (Request for Comments):https://www.rfc-editor.org/

These are detailed documents that specify the protocols and policies of the internet, including TCP/IP standards. They are available through the IETF RFC Pages. Internet Society:https://www.internetsociety.org/

A global, nonprofit organization dedicated to ensuring the open development, evolution, and use of the internet. It offers various publications and resources on internet protocols at Internet Society. Educational Institutions:https://ocw.mit.edu/

Many universities and technical institutes provide comprehensive notes and lectures online. Resources like MIT's OpenCourseWare or Stanford's Engineering Everywhere often have relevant materials on networking and protocols. By exploring these resources, you can delve deeper into how TCP/IP works, its role in modern communications, and the technical details underlying internet networking."

I copied the urls of the links because I dont know how to hyperlink on HN and the Stanford Engineering Everywhere link was broken. This is from the standard GPT4 ChatGPT web client. No plugins or browsing features enabled.

Re:The New York Times is suing OpenAI and Microsoft for copy... by Jensson on Wednesday December 27, 2023 @04:17PM (#38786849)
It should link to one of the articles about TCP it used as a reference to write that info blurb, not the TCP spec.

The problem is that those links doesn't link to where it got that text, it links to whatever that text linked to. Saying it is giving links is like saying that when I copy paste an article with links I am providing links to the source. No I am not, I am plagiarizing including plagiarizing those links.

So, it has read some TCP tutorials and wrote that blurb based on those. Don't you think it is fair that it links one of those to give credit? LLMs aren't capable of writing tutorials based on specs, they write tutorials based on tutorials it has seen, it should link to those.

Re:The New York Times is suing OpenAI and Microsoft for copy... by fennecbutt on Thursday December 28, 2023 @12:56PM (#38796224)
Presumably it can't link them, because it's been train on the data, not built on top of it. Gpt model doesn't include the sum of all training data, that's not how machine learning works at all (and overfitting on such a large and diverse dataset would be a monumental fuck up)
Re:The New York Times is suing OpenAI and Microsoft for copy... by abenga on Wednesday December 27, 2023 @02:22PM (#38785511)
I would be more impressed if it returned links to the specific RFCs and more specific pages elsewhere. What's a top-level link to OCW worth here? OCW is amazing, but has classes on practically everything. These are practically just domain names for "places to learn about the internet".
Re:The New York Times is suing OpenAI and Microsoft for copy... by MeImCounting on Wednesday December 27, 2023 @02:36PM (#38785653)
Well I asked it about tcp/ip generally and it provided general resources. Based on the context of my question thats about what one would expect. Its not perfect but it definitely can give urls to specific resources. It would be great if it got better at giving more specific links sure and some domains it can give more specific links than others for instance some git projects it can give precise references to docs while it doesnt seem to have the URLs for more specific courses on OCW, its not perfect but it is still a capability that it has.
Re:The New York Times is suing OpenAI and Microsoft for copy... by cjbprime on Wednesday December 27, 2023 @03:51PM (#38786600)
These are not citations. The point is that it does not / can not reliably cite the actual sources it used to prepare an answer.
Re:The New York Times is suing OpenAI and Microsoft for copy... by NemoNobody on Friday December 29, 2023 @03:17AM (#38802747)
Yeah ok, cause APA or some academic style is the to cite something professionally.

I'll be sure to tell everyone that uses the Internet

Re:The New York Times is suing OpenAI and Microsoft for copy... by zer0tonin on Wednesday December 27, 2023 @02:19PM (#38785469)
Even a middle schooler would be able to link the actual RFC 793 instead of just rfc-editor.org
Re:The New York Times is suing OpenAI and Microsoft for copy... by logicchains on Wednesday December 27, 2023 @02:21PM (#38785497)
From memory?
Re:The New York Times is suing OpenAI and Microsoft for copy... by tremon on Wednesday December 27, 2023 @02:29PM (#38785573)
No, from storage.
Re:The New York Times is suing OpenAI and Microsoft for copy... by jauntywundrkind on Wednesday December 27, 2023 @02:31PM (#38785594)
The ability to cite some rfcs is, to me, vastly different from being able to link to sources.

By far the wildest piece of this stuff is that it near completely obliterates any traces of where the outputs come from. The black box is trained, and yes sometimes some salient data pole rfc's are captures, but generally where each training comes from is not stored. That would largely defeat the purpose, would make the data it's crunching essentially incompressible, to store so much origin information.

Deeply unimpressed by this answer. This isn't linking it's sources, of where this response was trained upon. It probably got the write up & links from hundreds of other places.