Subscribe
Sign in
Claude by Anthropic
p/claudeA family of foundational AI models
Start new thread
Mat Sherman

Why Am I Not Loving Claude?

by•

I am not a developer. I am a sales/business guy who gets an obscene amount of value out of ChatGPT. It remembers who I am, seems to give me better answers than Claude, and I have the impression that it’s a leading model, whereas Claude is top 10—maybe top five, but not top three. With that said, all I hear online is that Claude is better on all fronts. It’s better for coding, more ethical, and will last longer than ChatGPT. I even hear it has better models. If that’s the case, why do I feel like it pales in comparison to my ChatGPT experience? Why does my own experience not align with the broader appeal I see for Claude every day? Is it because I don't code or am I just using it poorly?

Add a comment

Replies

Best
Burt Herman
I find that Claude's projects, under a paid plan, make it incredibly useful -- are you using them? If you upload some documents relevant to what you're working on, it does a great job referencing them, taking it into account when writing new text, etc. I've created personal projects with my resume and blog posts, work projects with text from websites, and more.
steve beyatte
@mat_sherman what sorts of stuff do you feel like it doesn't do better? For me @Claude by Anthropic is better at writing. That's out of the box, no special prompting "write me a blog post", "write me an email", etc. type of stuff. It does a better job of not sounding artificial and not sounding like AI. They also had Projects first so I added a good amount of external documentation in various projects that it does a good job of context-handling for. I know @ChatGPT by OpenAI does this now, too but because Claude did it first, I tend to use it more. Sonnet is the default coding model such that it's the out of the box default for @The Windsurf Editor and @Cursor . It's the best at this IMO. What else do people like/not like about Claude?
Mat Sherman
@claude @chatgpt @the @cursor @steveb So I think that my main issue is that whatever I ask Claude to do, I ask ChatGPT to do and it's 10% better. Not 25%, but a solid 10%. It goes into slightly more detail. Or is there a world where my mind has just drunken the ChatGPT KoolAid and it's actually the same quality and i'm just very biased? Also for a while, it seemed like ChatGPT just had more capability. Like crawling the web or a site. I know Claude will eventually have that stuff, but I am an impatient early adopter. I don't know, but I'm going to try Claude again and repot back results in this thread. I would prefer to use Claude, but i need to feel like it's better.
steve beyatte
@mat_sherman The ChatGPT team has built more and much faster. In a world where the models get commoditized (which seems like a given), then the original cry of 'AI wrappers have no moat' is going to be proven wrong exactly because of your sentiments.
Elena Tsemirava
For some tasks (like working with numbers, formulas, and calculations), I use ChatGPT. For writing texts, I prefer Claude. Maybe it's just a matter of personal preference.
Mat Sherman
@elenat yeah that makes sense. at the end of the day, I feel like all of these models will end up pretty similar, echoing what @steveb said above.
Gabe Perez
I explicitly use @Claude by Anthropic for writing or written content. I think it produces better output when trying to do more human linguistic challenges but falls short when it comes to thought process or logic compared to ChatGPT. I also wonder if we're so used to using @ChatGPT by OpenAI that we're actually prompting Claude incorrectly (it's a different AGI after all 🥴). I'm curious how are you using Claude and how much does it differ than ChatGPT? I'm assuming because it's sales and marketing there might be some planning or logic-based things that aren't quite up to your standard.
Kay Kwak
Launching soon!
It seems that Claude is better at finding and organizing information. And sometimes, GPT stops in the middle of a conversation. But as a pioneer, GPT has brought us tremendous change. That's why we're so familiar with it and can’t let it go.
Rachel Chan
Great share! I’ve also noticed that, although most people around me prefer Claude or other AI-generated products, I’m just more used to ChatGPT and can’t quit it, never really asked myself why😂. Never thought about the differences in prompts.
Emily Willis
Launching soon!

Hi Mat, great discussion topic! I agree with you. I've been using Claude for writing for months. Originally I found it way better than ChatGPT (at crafting copy that sounds more original, more unique, less robotic etc). I was drawn to it for its ethics & values also.


But it's really limited if you're producing long form content (eg 25K words instead of just 2K word blogs). The chats cut you off once you hit your daily limit. This happens to me often mid afternoon, when I'm at the last 2% of finessing a long form article & I'm locked out for the day - aaargh! This does my head in. Claude reminds me to "start a new chat" but when I do that, it doesn't remember all the learnings, I have to summarise them (the following day) & then start again in a new chat & it never remembers long form writing editorial prompt learnings - obviously not efficient.


The other problem is Claude can't search the web to source research & cite references properly - its referencing is woeful. They're dated & most are hallucinations or hybrid mumbo jumbo so you can't use them.


I've now started using ChatGPT which (using the editorial prompts crafted in Claude) is working well. The project folders are brilliant, it can search both uploaded resources/files as well as forums, social media, the web broadly, the citations are reliable & the tone of voice gets better the more you work with it & craft detailed editorial guidelins & tone of voice docs etc.


You can use it for research, for reasoning, critical thinking, writing, strategy, mind mapping etc etc.


Based on CLaude's recent data insights report, it's mostly used by developers & based on their latest email (received overnight) - their focus is shifting to developers - is that your understanding also?

Shabnam Katoch
After reading all the comments defending Claude I am definitely trying it out
Paweł Czajkowski
@mat_sherman I believe that you already play a bit with defining your own "style". If you haven't try it. I You can set rules for AI to generate responses which fit your expectations.
Luke

I have had a similar experience, at times I find that ChatGPT provides more relevant answers and can give information based on past chats while with Claude its like one has to 'jog its memory'. Although for more technical work like coding or analysis Claude has been extremely valuable for me.

Rephrasing things and asking the AI model what is missing/what it doesn't understand can be helpful to get more use out of both Claude/ChatGPT.