Meet Claude 3.5 Sonnet: First Impression of a model Superior to GPT-4o
ฝัง
- เผยแพร่เมื่อ 26 มิ.ย. 2024
- First look at Claude 3.5 Sonnet from Anthropic.
LINKS:
claude.ai/
Blogpost: www.anthropic.com/news/claude...
Timeline: tinyurl.com/3t93abjn
Model Card: tinyurl.com/25p7a4fj
API Docs: tinyurl.com/bde2zhw7
💻 RAG Beyond Basics Course:
prompt-s-site.thinkific.com/c...
Let's Connect:
🦾 Discord: / discord
☕ Buy me a Coffee: ko-fi.com/promptengineering
|🔴 Patreon: / promptengineering
💼Consulting: calendly.com/engineerprompt/c...
📧 Business Contact: engineerprompt@gmail.com
Become Member: tinyurl.com/y5h28s6h
💻 Pre-configured localGPT VM: bit.ly/localGPT (use Code: PromptEngineering for 50% off).
Signup for Newsletter, localgpt:
tally.so/r/3y9bb0
00:00 Introducing the Cloud 3.5 Sonnet: A GPT-4 Killer?
00:48 Benchmark Performance: Cloud 3.5 Sonnet vs. GPT-4
01:46 Cost and Availability of Cloud 3.5 Sonnet
02:59 Artifacts: A New Way to Interact with Cloud
05:24 Hands-On Demo: Creating a Flappy Bird Game
07:12 Image Understanding Capabilities
08:47 Conclusion and Future Prospects
All Interesting Videos:
Everything LangChain: • LangChain
Everything LLM: • Large Language Models
Everything Midjourney: • MidJourney Tutorials
AI Image Generation: • AI Image Generation Tu... - วิทยาศาสตร์และเทคโนโลยี
"Artifacts" is really the way to go. Keep a partition between the discussion and the created result. No longer a serial record with output copied over and over and over again.
ps. You are in my top tier of AI content creators. Practical, calm, intelligent.
Thank you for your kind words. I agree, it's a really useful concept. Have been playing around with it for a project and have found it to be more useful than code interpreter in chatgpt
Same here. It is tedious to every time, copy source, test it ,tell the llm (chat-gpt4 payed version) what it needs to change, ... . And go over the whole process again.
From my experience, even Opus was better than GPT4o. However the only issues with Claude were:
a) Cutoff date (more than a year ago)
b) Speed (slower in comparison)
c) UI/UX (e.g. inability to stop or change your previous query)
With all of these fixed in the Sonnet 3.5, I find everything better in Claude. Also coming from a company that I appreciate more in their attitude towards safety & transparency.
Surprise! 2 of the companies sharing the same base model. 2 faces of a 1 coin and owned by the same people.
@@Quitcool Spot on. Anthropic was founded in 2021 by seven former employees of OpenAI.
@@john_blues what's not spot on, is that these separate companies don't hold all else static
I still insist that 4o is a near-paradigm shift and is currently being handicapped. That ms speed over multi-modality/no handoffs is by far the greatest claim. Something underlying is quite different here.
Its just such a shame everytime i touch Claude even with a pro account the amount of messages I'm allowed to send is farrrrr too little and makes it hardly usable at all.
Just ran into this. I was so happy and ready to convert from OpenAI but that limit is ridiculous. At least give me a plan to pay more
This is why I just cancelled my pro account. Redicelously limited number of messages. So how good is their AI is irrelevant. They need to focus on their broken greedy business model!
I just use two Claude accounts and canceled OpenAI.
@@jaysonp9426 2 free accounts? But that means you have to start all over again and you lose the chat.
@@jaysonp9426 yeah I think I'll do that as well. Claude sonnet si far superior tbh
Claude 3.5 Sonnet looks amazing!
I agree
Incredible
this is my go-to site for new AI features, thanks
thank you!
Wow www another flappy bird demo, riveting!
🔥
Isn't it funny how they do their versioning? Now, Claude-3.5 isn't just better than GPT-3.5, but also surpasses GPT-4. So, they're basically saying that whatever the latest version of OpenAI GPT is (let's say N), their N-0.5 version will surpass it. This will continue unless OpenAI stops releasing updates until Anthropic's versioning catches up! :D
Wtf, you did not listen a thing to what he said, did you
It's another company which released their 3.5 model, not OpenAI
anything gpt-N can do, claude-(n-0.5) can do better XD
artifacts does not seem to be working for me. I entered the same prompt and got “I apologize, but I do not have the ability to generate edit, manipulate, or produce images.” I’m on the professional plan. Any advice?
I don’t understand why suddenly all these “models” are getting released AFTER ChatGPT came into existence? What were they doing before that?
Go to any youtube video that discusses unprofitable ideas...things that would hurt the oligarchy. You'll find tons of bots. The best example are videos discussing student loan debt forgiveness...something ppl would have to search for. Under these videos are a flood of comments from mostly bots ridiculing the idea of debt forgiveness. People who don't have student loan debt aren't searching for such topics...yet, we're supposed to believe that many of the 100s of the commenters are searchingfor such topics. In other words, the tech is used for profit and social manipulating purposes.
Is there an "artifact" history? If not, I see a situation where I had it generate something that I didn't save and would like to go back and get it. I can do that when the output and input are mixed but if there isn't a accessible history I won't be able to. Thank you for the informative video. Good stuff, as usual from you.
Yep, there is! Just tried it myself and it's similar to the way you can go back to earlier answers after editing a prompt in ChatGPT.
Yes sir! I wondered that too then I tried it myself and I can confirm the history is saved.
still no agents or web search?
in really it kills harder than in benchmarks
best AI content on YT, thank you for always informing us!
I wish videos would stop using the phrases 'Killer', 'Changes everything' and the like.
Wait but $3 per minute/per token?
$3 per Million input tokens :)
🎯 Key points for quick navigation:
00:00 *Cloud 3.5 Sonnet surpasses GPT-4.*
01:19 *Costs $3 per million input tokens, $15 per million output tokens.*
03:09 *Artifacts feature enhances user interaction.*
Made with HARPA AI
By YouSum Live
00:00:10 Cloud 3.5 Sonet: State-of-the-art model surpassing previous benchmarks
00:00:23 Artifacts feature: Enhances user interaction with Cloud
00:01:00 Performance superiority: Beats GPT 4 and previous models
00:01:30 Model availability: Accessible through API for immediate use
00:02:42 Image understanding: Enhanced capabilities demonstrated
00:03:00 Artifacts demonstration: Interact with model-generated content
00:04:20 Future model releases: Teasing upcoming Cloud versions
By YouSum Live
Can we use it for free on Amazon Bedrock?
its available there but not sure if you can use it for free.
@@engineerprompt OK I've checked AWS and no it's not free to use, however it's much cheaper than Claude 3 Opus which is great
Claude 3.5 Sonnet:
Price per 1,000 input tokens: $0.003
Price per 1,000 output tokens: $0.015
CloseAI Altman was trying to hide GPT-5 from us.
He even got a NSA director on the board.
But now he will be forced to release GPT-5. 😂😂😂
Claude is closed source as well.
@@john_blues they never claimed otherwise.
Plus they have some principles and a freaking Constitution.
That's way better than everyone else, including Elon
no way gemini is that high...
Gemini 1.5 Pro is severely underrated, Google aistudio is the way to use it with its full capabilities.
censored to the nth degree
that is expected from Anthropic models.
@@engineerprompt it is frustrating when you ask marginal questions such as write a PS script to set up new workstations and turn off MS defender in prep for auto deploying alternate AV. It refused to do this as if there is an individual rule on it. Makes it useless if it cannot simply generate the task. It has no issue with any other AV product. Just defender. Hard to trust its reliability when that kind of goal keeping is in play.
@@MrSCAATyou can set up the prompt for manual editing so most of the busy work is done