Will Claude 3.5 Sonnet with Artifacts impress you?!!!
HTML-код
- Опубликовано: 5 июл 2024
- 🔗 Links 🔗
Release - www.anthropic.com/news/claude...
Try here - claude.ai/
This video explores the benefits of Claude 3.5 Sonnet. The presenter discusses the different Claude models and their cost, and how Claude 3.5 Sonnet stacks up against ChatGPT. The video then explains a new feature called artifacts and how this feature is revolutionizing how people use AI.
Timestamps:
0:00 Intro and Overview
0:12 Claude 3.5 Sonnet Explained
0:43 Claude Models and Cost
1:07 Claude 3.5 Sonnet vs. ChatGPT and the Hype
1:34 Benchmarks and Performance
2:08 Artifacts and Code Interpreter
2:49 OpenAI Code Interpreter
3:01 Claude 3.5 Sonnet Artifacts
3:49 Artifacts Explained
4:03 How to Enable Artifacts on Claude
4:56 Simple SVG Demo with Artifacts
5:58 Simple Dashboard Demo with Artifacts
6:02 Human Face SVG Demo with Artifacts
6:47 Simple Dashboard Demo with Artifacts
8:21 Outro
❤️ If you want to support the channel ❤️
Support here:
Patreon - / 1littlecoder
Ko-Fi - ko-fi.com/1littlecoder
🧭 Follow me on 🧭
Twitter - / 1littlecoder
Linkedin - / amrrs 
Наука
![Eminem - The Death of Slim Shady [Album Trailer]](http://i.ytimg.com/vi/X0HIrS6kUYI/mqdefault.jpg)
Sonnet is exceptional; its answers are incredibly helpful and thorough. It's fantastic for children to learn and visualise.
It really is!
Impressive features. Thanks for the review.
meanwhile, openai not even shipping voicemode.
Sonnet is really good so far, pleasantly surprised how good it is
Anthropic really cooked something very good
Sonnet is good at coding, but the rate limits for paid accounts are still a bummer.
Sonnet 3.5 is super impressive and makes GPT-4, both with and without the "o," look bad, especially for coding. Its vision capabilities are way better than GPT-4.0. I can't even imagine how awesome Opus 3.5 must be. The only downside is the pricing, which isn't terrible, but if it were cheaper, I could hit the API every 3 seconds, all day, every day, to make my version of a real-time "Microsoft recall." Local options just aren't fast enough, and if I want to use Groq, their Llama 3 only supports an 8K context, while Claude supports 200K. It would be amazing if local inference could be real-time.
just tested drawing of a system architecture diagram. It used Mermaid diagrams. It didn't render because it couldn't handle a syntax error, but when given the error message manually, it fixed it. I'm impressed. However, it doesn't provide RAG, so I can't cancel my ChatGPT subscription, yet.
I like the Yet you added at the end 😁
@@1littlecoder yes, because it's definitely a huge step forward. After "agentic" the interaction now becomes more programmatic. I even don't have the word. It's like "manifesting" intermediate steps to the goal, while in the beginning with LLM text generation it's all "fluid". I think one of the next steps we all need is "generation" within boundaries. Imagine like give an XML schema and let it generate sythetic data, but based on the XSD.
yes - I am impressed - definitly step forward :)
I see you added graphics to create transitions. For it to work better I think there should be a sound (usually is some sort of sweeping/windy sound) to accompany the graphics. Ciao
Thank you. If I may ask a favor, could you point out a time stamp where the sound effect could be helpful?
@@1littlecoder I was refering to 6:55, when the graphic appears. Looked to me that you were using the graphic to transition to the next scene. If that is so, the standard would be to have a sweeping sound (it comes from the idea of a turning page, but nowadays any sound that gaves sense of transition can work, like a a swhoosh, or even a digital sound effect.. you have to see what you like best). I was trying to find examples, as is quite common, but I absurdly I can find any, when I do I post it here.
@@DCinzi Thank you so much for being so kind to get back to me with the timestamp. I'll look into it
this is absolutely crazy
Anything in particular?
Certainly impressed me… I did it before, but a lot more copy & paste was involved.
Thanks for putting this together. Can we see the prompts you used in the video?
Thanks. If it's hard to get from the video, I can put it on a word doc and share!
@@1littlecoder If you could do that, it would be very helpful. Thanks, once again.
how do i try this? claude 3.5 is available for free?
Yes it's available in free plan. The link in the description
Just built something with it at the UC Berkeley AI hackathon. Anyone go?
What did you build ?
The limitations are far too bad - basically even with pro you might hit the limit after 10 messages if your messages are too big - this is completely rubbish.
Also, teams plan is not clear
At least with ChatGPT you have a certain number of messages regardless of the length and load on their servers - when I pay for something I expect to be able to use it all the time - not when their servers are free!!
Is it free to use?😊
@@Jamielookscool yes with limited number or responses
Asak, Abdul Majid your giving very valuable knowledge. Are you available on LinkedIn? Tried searching you but can't find.
Va alaikum salam. I'm available here www.linkedin.com/mwlite/profile/in/amrrs
I ditched OpenAI today.
For Claude?
@@1littlecoder yes.
I have a sneaky suspicion, you are based in australia..
🤣 The Koala who Could
Wait for the s***ware coming to playstore and apple 😢
Switching from gpt to claude is nice and depressing at the same time
Why depressing 😁
@@1littlecoder well i often hit the message cap with gpt 4 than 4o came and i didn't hit it once. No switching to claude and it feels like i'm hitting it after 1 hour of coding
React developers ded
Should I have named the video like this ? 😂
Yann LeCun got a surprise beatdown in the hood for dissing LLMs! 🤖💥🏙