Hacker News Reader

Comment by MarvinYork | Hacker News Reader

This is a good example for what ChatGPT does: It makes shit up and presents it in a compelling format. You should always keep this in mind when trying to get information on a subject you are not familiar with yourself.

I use ChatGPT daily, but always specify the information I want formatted. (In bullet points, a table or for example a news article.)

View on HN

Can you please share what you use it for daily?

I use it for first drafts of some less demanding writing work. My workflow is more or less research with Bing Chat for summarisation and key point extraction, then feeding the results to GPT-4 so it uses the research as context for a first draft. Then the longest phase is editing, fact-checking, adding new information as required, etc. I’ve also created vector indexes for large documentation sites I work with and used langchain so I can query them, get sources, and have ChatGPT write summaries. Honestly, it doesn’t save a huge amount of time because then research phase was always the most time consuming to begin with: I still have to read and understand the material.

Of course, that’s only possible if the context content exists in the first place. If not, then it’s back to the old-fashioned technique of figuring things out for myself and writing about them.

I'm very interested in this work flow. Would you mind sharing the ChatGPT prompts that you use? Thanks.

I use it daily, sort of like a professional site/forum where someone immediately answers any question you may have. It may not be 100 % accurate, just like real people, but it's instant and personalized. (And in my experience GPT-4 is just spot on for anything a beginner might ask.) It's like training wheels for anything.

How do I set up some backup infrastructure for my home server? How does Btrfs handle read errors? How do I cook X if I don't have ingredient Y? What's the correct syntax for this command? How to analyse and interpret these measurements from this vermicomposting experiment (helping my girlfriend with statistics)? How do noise cancelling headphones REALLY work? Is it okay to have my samba.conf set like this (it wasn't)?

This is what I'm most curious about with AI. How often do you catch it being incorrect? The questions you're asking a fairly inconsequential if wrong, but how often do you fact check what it says?

You have to tell it. "Please be correct" it usually works

What was the accuracy and the difference between googling the question ?

I can't say anything specific about the accuracy, other than that it's good enough, but the difference between asking GPT-4 and googling the same question is night and day. And that's comparing it to Kagi, with my own search filters with boosted wikipedia, stackoverflow and generally scientific sources. Do you know the feeling that you want to ask a specific question about a specific, technical aspect of something, and you wish you could just tell the search engine somehow that you're just not interested in hand-wavy popular science average Joe half-incorrect answers? And if you actually find the one good answer on page 14 of some god-forgotten thread from 2006, the question differs in some tiny but important detail from what you're after?

Well that feeling doesn't exist with GPT-4. So far we've always been able to come up with something, together. If you don't like the first answer, you ask more questions. You can dig deeper. You can tell it what your guess is, what your intuition tells you about the problem, where exactly your uncertainty lies, e.g.:

  You: I don't understand phenomenon X.
  GPT: Oh that's easy, X is just [parrots wikipedia].
  You: That's fine, but I don't understand how X differs from Y, they sound like the same phenomena.
  GPT: I see why you'd think that, but Y and X actually differ in this detail called Foo that makes all the difference.
  You: I still don't get it, compare Foo to something similar that I know from normal life.
  GPT: Okay, so Foo is like an elephant who's too large to drive a car.
  You: Ooh, I get it now.

But here's what happens when I try that :

  Me : I don't understand this publicly documented AWS service with plenty of OSS examples, can you suggest how I would solve some edge case X
  GPT : Oh that's easy, invents bullshit that sounds exactly like what I need 
  Me : Googles the shit the GPT came up with and can't find any references
  Me : Maybe there is something similar - spends more time searching
  Me : GPT you're wrong - your solution doesn't exist 
  GPT : Oh sorry, here's the correct solution, comes up with more bulshit
  Me : Googles the shit that GPT came up with and can't find any references
  Me : Starts googling and solving the problem on my own
  Me : Finds out X can't be done with AWS service

I've had this flow many times now, it never resulted in valuable output from GPT vs just Googling since I need to verify everything anyway.

  Me : Review and come up with improvements to this code : CP code
  GPT : Suggest bullshit improvements, gives factually wrong reasoning, makes code worse
  Me : Points out everything wrong 
  GPT : Suggest other bullshit improvements
  Me : Points out everything wrong
  GPT : Returns basically identical code to input

  Me : Propose a design to this technical problem 
  GPT : Implements the solution in the most roundabout way and suboptimal way

IMO if I'm not sure what the output should be GPT is less than worthless it's actually convincingly misleading.

Is that with 3.5 or with 4? 4 is usually not afraid to tell it doesn’t know.

I've upgraded to 4 ASAP, all those examples are with 4. I tried 3.5 buy it was completely useless, 4 actually started giving sane output but still lying and hard to verify.

My only use case so far is stuff like "translate this model to openapi schema" because it was faster than setting up the tool, and similar tasks that are easy to verify and boring to type.

It's great at little tedious tasks, as long as you review its work.

Like converting a code sample from one language to another, or one format to another (e.g. raw css to object syntax), or complex replaces where the regex gets tricky- also great at writing simple docs.

Also use it for micro tutorials if it's on a topic older than a couple years old.

Or converting unstructured (or semi-structured) text to structured format. My most recent example with GPT 3.5: copy-pasting a piece of a website with a list of dates and events, and prepending it with "Please rewrite it into ICAL format".

https://cloud.typingmind.com/share/dc9b6a87-cd00-43e9-a7ea-6...

The result I then could paste into a text editor, save as .ical, and import straight to my calendar.

Not daily, but often: questions about history. GPT3/4 demonstrates good historical insight and the ability to make nuanced arguments. For instance, I like the mention of some of Reagan's "liberal" policies in this answer:

"Can Ronald Reagan be considered a liberal?

Ronald Reagan is generally considered a conservative politician. He was a member of the Republican Party and served as the 40th President of the United States from 1981 to 1989.

Reagan's political philosophy was based on a belief in limited government, individual freedom, and free markets. He advocated for a reduction in government regulation and taxes, increased defense spending, and a strong national defense. He was also a strong supporter of traditional social values, including opposition to abortion and same-sex marriage.

While Reagan did take some positions that could be considered more liberal, such as his support for amnesty for undocumented immigrants and his advocacy for the abolition of nuclear weapons, overall his policies and beliefs aligned more with conservative ideology. Therefore, he is not typically considered a liberal."

Now ask this to a human. You will get a polarized answer "No, Reagan is a conservative icon!!!!!!" lacking any nuance.

'Considered a liberal' by who? ChatGPT doesn't tell us. I only mention this because the academic definition of the term 'liberal' is very different from the colloquial understanding.

> Now ask this to a human. You will get a polarized answer "No, Reagan is a conservative icon!!!!!!" lacking any nuance.

Depends on the human. Ask a political science professor and the result will probably be pretty similar, plus an earful on the evolution of the two major US political parties' positions since the 1980s and how that complicates these kinds of retrospective judgements.

Proofreading, re-writing, code snippets, basic general information on a topic, boilerplate text, summarising. Makes a pretty good assistant.

Not OP but I am using it to finally shape my blog into what I always wanted. It still takes me up to 100 hours to finish an article (sometimes more than what I used to spend on an article before ChatGPT) but my articles are so much better and the research process is so much more insightful and fun.

GPT-4 literally writes software while I sleep.

Reply on news.ycombinator.com