Claude 3.5 Sonnet for sparking creativity
AI技術の進化が加速している中、Anthropic社が開発したClaude 3.5 Sonnetはその最先端を行く存在と言えます。この最新バージョンでは、Artifactsという新機能が追加され、ユーザーがClaudeとのやり取りをさらに深めることが可能になりました。
Artifactsを使用することで、Claudeに対してドキュメント、コード、mermaidダイアグラム、ベクターグラフィックス、さらにはシンプルなゲームの生成を依頼することができます。これらの作成物はチャットウィンドウの隣に表示され、リアルタイムで確認し、改良し、作業を進めることができます。
さらに詳細な情報はAnthropic社の公式ニュースページでご確認いただけます：https://anthropic.com/news/claude-3-5-sonnet
Claudeをお試しいただく場合はこちらから：https://claude.ai
Claude 3.5 Sonnetの魅力は、AI技術を活用した創造性豊かなコラボレーション体験が可能となった点です。今後もAnthropic社の取り組みから目が離せません。
This is such a cute demo of the claude's capabilities. Well done to the anthropic team!
エグいな、もう何でもありだな
しかしプロモーションと現実の乖離はあるからどれだけ実用出来るかは、、、
Is anyone able to use Claude as a copilot to help stay focused and on task? Especially if you have ADHD. I LOVE Claude so far… it's such a gentle and ethical tool!
Tried it but didn´t work 🙁
awesome!
Gemini is the worst, ChatGPT somewhat ok but Sonnet is actually the king. I started using it just a couple of weeks ago and was mind-blowned how much of complex instructions it can handle without errors while ChatGPT just outright fails and Gemini doesn't even know what to do, with the exact same command. I'll get their pro version just to support their work. On social and political questions, Gemini spits propaganda straight from US state dept. while Sonnet gives actual honest factual answers too. Kudos to you guys.
After using it for a month, I think it is better than gpt4o！
Hello! I very like your product and often use it. Will it ever be possible to create PowerPoint presentations? It might be very useful.
Amazing stuff
I tried using it as a dev co-pilot and could not believe how fast I ran out of messages. And I even paid for the Pro plan, which just gives you more messages. It's a ridiculously low limit to start with. But whats worse is Claude's short-term memory loss. It continuously forgot my instructions, or simply ignored them. I had to remind it countless times and regenerate which really burns through messages fast. And it would edit parts of the script that I did not ask it to edit. No matter how many times I reminded, in an effort to correct the behavior, it continued to repeat the same mistakes. Claude is also way too fast to offer substitute code or an alternate approach when something broke or didn't run as expected, rather than try to figure out what went wrong. This tendency will send you so far down a rabbit hole, breaking things that once worked along the way, that you have to start all over again with a version that did work from twenty prompts previous. All of this makes the limited number of messages for a Pro subscription unusable. You'll get more out of the free version of ChatGPT.
It's also worth mentioning that Anthropic provides zero support to paying Pro subscribers. I've sent four emails over one week and haven't received a single reply.
This does not work in the workbench! LAME!
It needs image generation better one then I'll definitely switch from closed ai
Yeah much respect and thanks ❤️❤️❤️❤️❤️❤️❤️
tried this prompt and apparently no – "I apologize, but I'm not able to create, generate, edit, or produce images, including 8-bit style graphics. I can only perceive and analyze existing images." – is this because I'm on the free version or a country thing?
excited for the 3.6 release where "Claw'd" makes it giggle too
People are watching this not realizing how hard it was for bots just a year ago. This is amazing.
This just accomplished a game I've been working on for the past two months.
i type in the exact same commands and it did not turn out to be the same result. mine is not playable
Amazing features, but why you exclude Brazil yet? Why you hate us? I wanted to use Claude since the start and still waiting…
Works only with React?
Claude Prime minister! There is still time to be elected in France, it'd be a bit of hope
😲
My name is Sam and I was like😲
They want a phone number to register / spam….(Ps the free phone numbers don't work)…I'll have to pass for now, I get enough spam on my phone…
I just finished running some basic coding experiments with Claude vs ChatGPT … . Claude output is way nicer, easier to read to the point and fast. But its limited .. I can work with ChatGPT all day, throw thousands of lines of code at it. With Claude, it complained after 10 minutes of usage and fraction of the code. It told me I was 10% over my limit and wanted to start a new conversation .. So its really only good for short tasks? This is with a paid subscription .. Meh – For my use case seems ChatGPT is much better, going to cancel until it changes limits.
how ce the ai doesn't want :create an .svg version of this logo.
I subscribed today. You made me to leave OpenAI with all of my projects. Next week I’ll change all scripts as well.
I love the way Claude communicates. It never discriminates me, because I tried to engineer a prompt that helps to identify certain types of inputs.
It was frustrating with ChatGPT. It always tried to discuss the content, instead of helping me with the engineering process.
I’m glad that I tested Claude! And I feel rescued from ChatGPT. ❤
I was tired sam Altman propaganda, this is real Ai
why Sudan is not among the countries which you can use their phone numbers for verifications?
Amazing job! I've been using Claude for several months now. It's much better than the CloseAI
Ohhhhh dang. I don’t think I was using it right
Hey.
When will there be a Android App?
Yes
I've had some amazing results with Claude 3.5 Sonnet but when I try to build a simple game to test, as per the example, I am told Claude currently is not able to run the code it generates – am I missing something? I tried a very similar game creator demo as shown in the video.
Sometimes this small demo could open so many possibilities to people eager to do something but no coding knowledge… I just want to say Thank you 🙏
It looks and sounds like something from space kangaroo! 🤣🤣Some inspiration was definitely taken from there!
Cool, but like all AIs nowadays, I bet it can't do blood or guns. May upset someone somehow, you know.
Love the "Good evening, Sam" at the end. No such thing as a coincidence.
Major W from Claude. I hope to see you push your models even further. Perhaps a new version of opus?
I love Claude 3.5 for this.