> slowly / painfully delete every single line one by one before rewriting
I think this one is a visual artifact. I think it always rewrites artifacts from scratch internally, but the UI tries to make it look like it isn’t doing that. The result is that it looks like it’s deleting lines from the inside of the text instead of just writing them.
That said, I’ve also experienced all of your complaints, except the bad edit one (but I see that all the time in Claude Code/Aider).
For the bad edit one, to clarify, I mean, it would "put its edit pointer" at the top of the file and start deleting/inserting there, instead of where it "intended to" which would have been somewhere in the middle, for example.
---
Erroring in that way specifically makes me think the "deleting every line" issue isn't just a visual artifact but ¯\_(ツ)_/¯
I tell Claude not to put anything in the artifacts because it always fucks then up. I see it writing code in it and then it disappears leaving the last artifact in place causing me to waste tokens having it recreate the artifact.
I find it useless most of them time and wish I could disable it.
The extra system prompt can definitely cause some performance issues, and it can over use them. The deleting every line behavior is gone though. It's definitely not something you should turn on for every conversation, but it's quite compelling for creating little capsule web apps.
I’m somewhat indifferent to artifacts. While they can produce useful code occasionally when the codebase is relatively small, they become excessively cumbersome once it reaches around 500 lines. This makes it difficult to even add or remove code. If you want to change a few colors in a webpage, they have to rewrite the whole artifact. It frustrates me that you can’t edit a single line in an artifact; you have to rely on the LLM to do it.
I have ran into a few situations where there's some console error that is fixed by rewriting the thing from scratch. Artifacts are awesome for being able to display some sort of interactive tutoring tool I've made on the fly, but realistically I could just have that HTML/CSS/JS file myself and push it to my own site relatively easy. For others though, it's hosted there easily
I think that's a good decision. They know their market and it's intended for small projects and demo mostly from non tech people. And they did not built half baked editor for which people would have further complaints about. AI assisted coding is a whole different thing and there are many players.
I totally missed you can embed AI features directly into claude artifacts. I ran the first GPT store hackathon in NYC and I feel like people were very underwhelmed. What could you do in a custom GPT you couldn't with a shared chat? The integration felt very annoying. Claude I feel nails how easy it is to build + how easy it is to share
I'm guessing the OP is looking at this through the lens of more mom and pop users on the chat interface rather than the greater product line of these companies with all their dev tooling.
Claude Code is radically better than anything ChatGPT has right now. If anything it's the other way around -- ChatGPT has the mom and pop mindshare and more first-mover name recognition, but they have an objectively worse product for 'serious' users.
yeah, it's been a bitt iffy for me this week. took a few days off and came back yesterday and in the mean time feels like claude code took 15 steps backwards
I tried to make an artifact that would simplify Wikipedia articles [0] but the artifacts stubbornly won't let you do ANY input into them, not even via query strings. I think I'd be able to make cooler artifacts once they allow more input/output stuff. I understand the security issues, and it makes sense to roll this out slowly, but I want it now!
How stable are these now?
I turned off artifacts months ago because it would:
- frequently update code incorrectly / bad edit diff
- act like it updated / created an artifact when it just did nothing
- slowly / painfully delete every single line one by one before rewriting
- use artifacts for things that shouldn't have had any code written at all
Just wasn't worth the value it provided. This was before claude code.
> slowly / painfully delete every single line one by one before rewriting
I think this one is a visual artifact. I think it always rewrites artifacts from scratch internally, but the UI tries to make it look like it isn’t doing that. The result is that it looks like it’s deleting lines from the inside of the text instead of just writing them.
That said, I’ve also experienced all of your complaints, except the bad edit one (but I see that all the time in Claude Code/Aider).
For the bad edit one, to clarify, I mean, it would "put its edit pointer" at the top of the file and start deleting/inserting there, instead of where it "intended to" which would have been somewhere in the middle, for example.
---
Erroring in that way specifically makes me think the "deleting every line" issue isn't just a visual artifact but ¯\_(ツ)_/¯
Technical term for "its edit pointer" is "the cursor".
I tell Claude not to put anything in the artifacts because it always fucks then up. I see it writing code in it and then it disappears leaving the last artifact in place causing me to waste tokens having it recreate the artifact.
I find it useless most of them time and wish I could disable it.
You can. Go to "Connectors" in settings and toggle it off.
The extra system prompt can definitely cause some performance issues, and it can over use them. The deleting every line behavior is gone though. It's definitely not something you should turn on for every conversation, but it's quite compelling for creating little capsule web apps.
I’m somewhat indifferent to artifacts. While they can produce useful code occasionally when the codebase is relatively small, they become excessively cumbersome once it reaches around 500 lines. This makes it difficult to even add or remove code. If you want to change a few colors in a webpage, they have to rewrite the whole artifact. It frustrates me that you can’t edit a single line in an artifact; you have to rely on the LLM to do it.
I have ran into a few situations where there's some console error that is fixed by rewriting the thing from scratch. Artifacts are awesome for being able to display some sort of interactive tutoring tool I've made on the fly, but realistically I could just have that HTML/CSS/JS file myself and push it to my own site relatively easy. For others though, it's hosted there easily
I think that's a good decision. They know their market and it's intended for small projects and demo mostly from non tech people. And they did not built half baked editor for which people would have further complaints about. AI assisted coding is a whole different thing and there are many players.
I totally missed you can embed AI features directly into claude artifacts. I ran the first GPT store hackathon in NYC and I feel like people were very underwhelmed. What could you do in a custom GPT you couldn't with a shared chat? The integration felt very annoying. Claude I feel nails how easy it is to build + how easy it is to share
Note that this is new; Artifacts launched almost a year ago, but the API to call Claude from within them was added just three weeks ago.
A great highlight on Claude’s new strategy. I agree authentication and payment are key features for user lock in.
"Claude is kicking ChatGPT's butt"
I'm guessing the OP is looking at this through the lens of more mom and pop users on the chat interface rather than the greater product line of these companies with all their dev tooling.
Claude Code is radically better than anything ChatGPT has right now. If anything it's the other way around -- ChatGPT has the mom and pop mindshare and more first-mover name recognition, but they have an objectively worse product for 'serious' users.
Have you used it this week? Quality has got really bad.
yeah, it's been a bitt iffy for me this week. took a few days off and came back yesterday and in the mean time feels like claude code took 15 steps backwards
Has anyone come across any really cool artifacts? I'd be curious to see
I tried to make an artifact that would simplify Wikipedia articles [0] but the artifacts stubbornly won't let you do ANY input into them, not even via query strings. I think I'd be able to make cooler artifacts once they allow more input/output stuff. I understand the security issues, and it makes sense to roll this out slowly, but I want it now!
[0] ended up making it a browser extension instead https://mattsayar.com/simple-wikiclaudia/
Simon Willison is an incessant champion of AI tinkering. This is a bit dated, but here's a post specifically on his Artifact builds: https://simonwillison.net/2024/Oct/21/claude-artifacts/
Here's all his posts tagged with claude-artifacts: https://simonwillison.net/tags/claude-artifacts/
Color palette generator: https://claude.ai/public/artifacts/719b00a3-66e7-46c7-b90d-a...
I like the use case for mini design exploration tools