Speaker A
Claude Opus 3.5 has been released. A new model from the company Anthropic. Usually, people ask neural networks for text, images, and so on. But in this video, we'll see things on a whole different level, tasks that used to
Explore 10 incredible examples of Claude Opus 3.5 AI solving complex tasks across fields like education, medicine, design, and business with ease.
Key Takeaways
- Claude Opus 3.5 significantly advances AI capabilities in diverse real-world applications.
- AI can create professional-grade outputs in design, education, and medical visualization with minimal input.
- Open-source projects and free guides enable broad accessibility to powerful AI tools.
- AI-driven solutions can disrupt traditional industries by automating complex tasks efficiently.
- Community engagement and iterative improvements are key to evolving AI applications.
What the video covers
- Claude Opus 3.5, a new AI model by Anthropic, performs complex tasks previously requiring teams, using simple prompts.
- Example 1: AI-designed 3D-printable bridge supporting 59 kg, outperforming other models like GPT-4o and Elon Musk’s Grok.
- Example 2: Creation of interactive educational websites explaining concepts like camera focusing, boosting the creator’s social media growth.
- Example 3: AI-built 3D viewer for dental jaw scans, enabling patients to visualize medical data without specialized software.
- Example 4: Generation of professional-quality motion design videos from minimal prompts, challenging traditional design industries.
- Example 5: AI-developed interactive apartment design tool with 2D/3D views and customizable features, released as open source.
- Example 6: Visualization of warehouse shipment tracking as a computer strategy game, inspiring real business applications.
- Additional examples include AI-assisted renovation planning, business digitization, Roblox game creation, and more.
- Bonus content includes free guides on using Claude Opus 5.5, animation creation, and beginner courses for ClaudeCode.
- The video showcases AI’s expanding role in creative, educational, medical, and business domains, highlighting practical use cases.
Chapters
- 00:00Introduction to Claude Opus 3.5 and AI capabilities
- 01:07Example 1: AI-designed 3D-printable bridge competition
- 02:11Example 2: Interactive educational website creation
- 04:27Example 3: AI dental 3D jaw scan viewer
- 05:29Example 4: Motion design video generation
- 06:46Free guide and animation creation with Claude Opus 5.5
- 07:53Example 5: AI apartment design tool
- 08:58Example 6: Business shipment tracking visualization
- 10:04Additional examples and community engagement
- 12:17Roblox game creation and closing remarks
Full Transcript — Download SRT & Markdown
Speaker A
take entire teams to complete, which are now available with a single prompt. Out of everything people have done with Opus on Twitter abroad, I’ve selected the 10 most spectacular examples for you. These cover various fields: education, health, home repair, and
Speaker A
business. So, surely some of this will be directly relevant to you. And for those who haven’t worked with Claude yet, there’s a bonus guide on how to get started from scratch. Let’s go.
Speaker A
We’ll start with a competition between neural networks. Do you know what it’s in? Bridge design. Five top models were given one task: design a 3D-printable bridge that can support the greatest possible weight. The conditions were: a 60 cm span, no more
Speaker A
than 500g of plastic, and no more than 18 hours of printing time. Well, basically, the neural networks did it all. And here is your result. The bridge from Kimi—that's a Chinese model—didn't even assemble. Elon Musk’s Grok bridge fell apart, and
Speaker A
that was without any load at all. Meta gave up at around 12 kg, and GPT-4o/Astra even earlier, at eight. But the bridge from Claude Opus 3.5 held 59 kg.
Speaker A
And those were just barbell plates. And at the end, a person even stepped on it. And it printed faster than the others. 9 hours versus almost a full day for Astra. By the way, the creator has a
Speaker A
whole "Agent Wars" series. And this was just the second episode. He has already announced the third one. So Kimi and the others still have a chance to redeem themselves, and we have a chance to watch this madness in future videos.
Speaker A
Example two. From the educational sphere. A person asked Opus to explain how a camera focuses. You know, one like this. But not with text, through some kind of interactive experience. In the end, he got this mind-blowing website. You can rotate everything,
Speaker A
look at everything, and perfectly understand how a lens actually focuses. Claude put this together in one go in 1 hour and 26 minutes. And if you convert it to tokens, this masterpiece cost 25 dollars and 66 cents. And the author
Speaker A
didn't stop at lenses. Now he has 29 such explainers on his site: how a car differential works, why ice cream has different consistencies, how Starlink connection works, and a whole lot more.
Speaker A
At the time of the first website, he had 500 followers on Twitter. Then it all went viral. The post got millions of views, and he grew his account to 9,000 in a couple of weeks. And the person, essentially, found a
Speaker A
mini-calling of his own—I don't know, visualizing different things with neural networks. Simply because he objectively does a great job. And a story like this suits teachers, tutors, and really anyone who explains things to others. Plus, it's just such
Speaker A
addictive content that's pleasant to watch and even play around with. All links are in the description. Next, medicine. Each of you has surely been to a dental clinic and seen them viewing 3D jaw scans on a big monitor
Speaker A
in some special program on a computer that looks like it's from the fifties. If you are given such a scan to take home, you won't be able to open it just like that. It’s not a Word document; it’s hundreds of files in a special
Speaker A
format that obviously require highly specialized medical software to view. So, before a root canal treatment, a person had a scan taken, and the dentist said something like: "You won't be able to open it yourself, we'll take a look in a week." Blah-blah. But the
Speaker A
patient decided to experiment. He asked Claude to make him a viewer for such files. One prompt. And there you have an interface. Jaw slices from three angles. A 3D model with teeth, roots, and a ruler you can use to measure bone.
Speaker A
And it feels, if not more professional, then at least better-looking than what the doctors end up using. And the guy is obviously not from a medical background at all.
Speaker A
It just popped into his head, and he went and did it. Yes. And yes, a neural network is an unrecognized medical specialist. I’m also 100% not a doctor. But in my opinion, if you show a neural network the slices from such a
Speaker A
viewer, it will definitely be able to tell you the basics. And you can go to the dentist for a diagnosis when you have a free slot. Our fourth example is about motion design. These are the animations you see every day. You don't
Speaker A
even think about them. Intros at the beginning of TV series. Commercials where logos and numbers fly onto the screen beautifully, animations in your phone apps, and so on. This is done by designers—a distinct profession. And it feels as though with the release of
Speaker A
the new Opus model, Anthropic has dealt irreparable damage to an entire industry. You'll understand why in a moment. Here is an example of a specific prompt from Twitter. Make a dynamic fifteen-second motion graphics video that a stunning motion designer
Speaker A
would show off. Give it your all. And to this seemingly very poorly described request, Claude produced a video that looks really stylish. It’s not some "neuro-slop," but actual motion design from someone with a taste for modern trends and so on. And at the end, a
Speaker A
signature: "Motion Designer," and in a separate font: "Open for new projects." Claude actually went and signed his work. It’s strange he hasn’t created an email address yet. Well, Claude is open to new projects, so let’s hire it, and do it for free. I
Speaker A
recorded a short video guide showing how to make this animation on OPC 5.5 without a Claude subscription, for zero dollars. You can find this video guide at the first link in the description, absolutely for free. There’s also a
Speaker A
pack with all the examples from this video. And if you don’t really need Monday, I’m throwing in a "Claude from scratch" course as a bonus, showing your first steps in ClaudeCode and how to build several ready-made projects. In short, you’ll definitely
Speaker A
find something interesting specifically for you. Now, this example is for everyone planning a renovation. A guy gave Opus his apartment layout and asked it to create a design tool. Here is what he got in an hour. In 2D, you
Speaker A
arrange furniture in real dimensions. It snaps to walls, you measure things, tear down non-load-bearing walls, change the flooring, and with one click, the apartment smoothly rises into 3D, and you can view it from above or walk through the rooms in first-person.
Speaker A
Ordinary floor planners can do similar things, by the way, but they are often paid software that clearly wasn’t made for you. But here, it’s just one image, an hour with an AI, and you can implement any functionality you want.
Speaker A
The author released the project code to the public. You can feed it to the AI along with your own layout and any plans for improving the software. Feel free to use it. The sixth example is the most liked on Twitter from the
Speaker A
entire selection. And it's about business. A guy just decided that tracking warehouse shipments shouldn't look like boring accounting in spreadsheets. Why not visualize it all in the format of a computer strategy game? You click on a shipment and see
Speaker A
whose cargo it is, where it’s coming from, and when it will arrive at the gate. And below is the entire order journey: confirmed, packed, loaded, in transit, and so on. It looks awesome.
Speaker A
For the author, it’s still just a concept—not a product with real data—but I came up with a really cool idea. Look, I have a friend named Edgar. He has a real business in Moscow, a pretty large vending machine chain with
Speaker A
50 units across the city. He’s a huge fan of AI. His entire business has long been digitized by agents. I suggested he do something like this for his machines. He really liked the idea. So I think we’re going to do it. I’ll
Speaker A
show the whole process in my Telegram channel step-by-step, from scratch to a finished prototype. There will be a series of posts, maybe we’ll even do a stream. So make sure to subscribe to the Telegr
Speaker A
Children's math. A Japanese author asked Claude to create an arithmetic trainer for kids who get bored. A child solves long-addition problems, and for every correct answer, they get a prize: confetti, banners, new little animals.
Speaker A
And each time, it looks more and more impressive on the screen. A sort of dopamine attraction. It was made in just an hour to test the model, and the code was released to the public. And do you know what I thought? AI is an
Speaker A
absolutely bombshell topic for parents. Let’s say you have a child who struggles with something like math, is bored, lazy, and so on. At the same time, they surely have favorite cartoons, video games, book characters, and so on. And you can simply build a
Speaker A
personal trainer for your child for these kinds of school subjects, built around their favorite stories. And your child will enjoy studying math or any other disliked subject. Just because it will have game mechanics built in, related to, I don't know, Minecraft and
Speaker A
so on. In general, definitely take note of this. Our eighth example is about music. I think many of you have seen the Lofi Girl channel on YouTube. If anyone doesn't know, it's a 24/7 stream where a drawn anime girl studies while
Speaker A
lo-fi plays in the background. I mean, that kind of calm music with a pleasant beat that many people love to work or study to. I happen to be one of them.
Speaker A
And this guy made a site called Lofi Cities using Opus 5.5, with pixel art of night cities and endless lo-fi on top. It has Paris, Dubai, Tokyo, Rio.
Speaker A
There are no CIS countries there, by the way, so think about it. And the most interesting thing is that the music here isn't a playlist. Every few minutes, a new track is composed right in your browser. Drums, bass, melody,
Speaker A
with vinyl hiss on top. And all of this is synthesized on the fly without a single recording. And this is just a very addictive site. Definitely go check it out and poke around. I’ve started playing it for myself now
Speaker A
because, well, it’s very atmospheric. The ninth example is about Roblox. If you think of Roblox just as a children's game, you're wrong; the maps , that is, the individual games inside Roblox, are made by regular people. And this is a niche with insane earnings.
Speaker A
Creators of popular maps earn tens and hundreds of thousands of dollars a month. And just over the last year, Roblox paid out about 1.7 billion dollars to creators. And you can connect Claude directly to this. It will draw the map itself, fix scripts,
Speaker A
build models, and launch the game for testing. So this is a very cool window of opportunity to make your own maps like this, release them to the public, and monetize them. And the example here , by the way, is quite simple. The
Speaker A
author of the tweet says their daughters love squishies. Well, those soft anti-stress toys. And, obviously, they also love Roblox. Before bed, he fed the game plan into Opus, and in the morning, he woke up to a world with
Speaker A
11,000 parts. Then they spent a week editing the map with the kids until, well, it became truly fun. And now the custom "Squishy Hunt" game is published in Roblox. You can go in and play it.
Speaker A
So, take a closer look at this niche. The next example is related to YouTube.
Speaker A
One guy made a video where all the animations and even the editing process itself were put together in Claude Opus . That video of his took off with 250,000 views. That guy, by the way, is me, and it was my previous video about
Speaker A
GPT-6 Astro. If you haven't seen it, definitely check it out. And all the animations you are watching right now were also made in Opus 5.5. We already broke down the entire process from start to finish with the guys. on a
Speaker A
stream in my private community Gara 403 , where we have a platform with video guides, articles on fine-tuning Claude, and much, much more. In any case, there is a free guide in Telegram on how to make cool animations, even without a
Speaker A
Claude subscription. But if you are interested in the details in the Garage , you are always welcome. And please, leave comments and likes, because, well , it pushes the video into recommendations. I understand that you like this kind of content, and I hurry
Speaker A
to get to work on the next episode. Seems like a win-win to me. Bye everyone.
Topics:Claude Opus 3.5Anthropic AIAI bridge designinteractive education3D medical visualizationmotion design AIapartment design AIbusiness visualizationAI tutorialsneural networks
