Future Stacked

AI, Tech and Robotics || Bringing the future to you. || DM for collaborations || 📩 - [email protected]

[Future Stacked @FutureStacked]

Pasta boiling over, one kid asking for help with math, me telling Nori to remind me about picture day.

Chaos. Normal chaos.

Then my phone lit up. It had sent us a turkey for Thanksgiving

I turned off the stove, sat on the kitchen floor, and just laughed.

It felt like someone walked into my messy kitchen, saw everything I was juggling, and said, "here, let me take one thing off your hands."

Go chat with Nori, maybe it'll take one off yours too

[Nori @Nori_FamilyAI]

Nori’s Thanksgiving Turkey Drop is live. 🦃
Chat with Nori about dinner, family schedules or holiday to-dos. Eligible chats give you a chance to score a randomly dropped $50 gift card.
Less planning. More time together.
See how it works: https://t.co/OAZxJ1ex2F

[Future Stacked @FutureStacked]

@gakkuai like the idea of AI helping kids think through problems instead of just giving them answers.

That's the kind of learning experience worth building.

[Future Stacked @FutureStacked]

Apple just dropped a new Apple Pay ad, and it shows how fast paying with your iPhone can be.

[Gakku.ai @gakkuai]

We’ve raised $51 million to rethink how kids learn with AI. Our bet: conversational learning.
Language isn’t just how we communicate. It’s how we think. That’s the idea behind Gakku, our AI learning coach. The learning process itself is the product—not something to automate away.
Does expression matter more in the age of AI? This time, the AI is asking.
Watch Raku, our AI learning companion, interview co-founder Annie Chen @anniechen419 , who created him.

[Future Stacked @FutureStacked]

Claude Fable 5.2 turned a sunset daydream into a world you can actually walk through.

[AI Frontliner @AIFrontliner]

🚨 BREAKING: FABLE 5.2 IS HERE!

But is it worth switching from Opus 5.5?

Here’s everything you need to know about its features, pricing, and limits:👇

[Future Stacked @FutureStacked]

@truefoundry London sounds like a great starting point.

Congrats to the team on this milestone!

[Future Stacked @FutureStacked]

@IFM_AI Being able to predict what happens before taking action could be a big deal for robotics.

Curious to see how well it works in the real world.

[Future Stacked @FutureStacked]

The claim in the PAN 2 release we most want to see tested: it handles combinations of conditions it never saw together in training.

That’s IFM’s claim, and we are curious how far it stretches.

[Institute of Foundation Models @IFM_AI]

Introducing PAN 2: an open foundation world model for physical intelligence.

A language model predicts the next word. PAN 2 predicts the next world: what happens after each action, grounded in time, physics and space.

Towards AI that simulates “what if?” before it acts.

More details: https://t.co/lUX7igPOiF

[Future Stacked @FutureStacked]

EU servers, US law.

Under the US CLOUD Act, a US provider can be required to hand over data even when it is stored in Europe.

That is why TrueFoundry's EMEA launch is a big deal, because the TrueFoundry AI Gateway runs in your own environment with access control, SSO and audit logging built in.

[TrueFoundry @truefoundry]

TrueFoundry is expanding across EMEA, with London as our regional hub 🇬🇧🇪🇺

[AI Frontliner @AIFrontliner]

🚨 BREAKING: ChatGPT now teaches how to BUILD PASSIVE INCOME from scratch. For free.

Here are 8 prompts to earn Money While You Sleep:

[Future Stacked @FutureStacked]

Your iPhone now shows up on the touchscreen MacBook Pro, and you can tap through it like the real thing.

[Future Stacked @FutureStacked]

Seele03 builds games straight into real engines like Unity and Godot, not just simple web demos.

When it's done, you can download the full project files and change anything you want, from animations to level graphics.

It also pays attention to things like combos and hit feedback, which is what makes a game actually fun to play.

[Grummz @Grummz]

Seele is a company that creates AI tools for game making. Recently they invited me to participate as they took on a 5-day challenge to push their new model, Seele03, to its limits and move closer and closer to AAA with a Wuxia-style action game.

The product consists of their AI model, asset generators, and target engines such as web (three.js), Unity, or even Unreal. What's nice is that their site has everything in one place, from mesh generators and rigging to sound fx and sprite work. They also provide common templates for most types of games. Having it all work together seamlessly saves a lot of time. You can get a game running very quickly, in 20 minutes or so. (20 minutes! I remember when this took months and months.)

So what can they do in 5 days?

Their progress video, below, shows them iterating through creating a Phantom Blade Zero-style game. This is not reverse engineering, but made from scratch. Over the course of development, they sent me video updates on their progress. The 03 model is designed for iterative development, and I gave them some advice as they took it from basically nothing to the final version of the video today. The progress the model made day to day was rapid, and the traces actually factored in something that I don't see that often: game design.

Most models focus on the graphics when making a game, but Seele03 is thinking about combos, hit feedback, everything that really matters for how a game feels and plays. One day, I think that's going to be the core of what makes AI game making viable.

After you are done prompting and refining your game, it's not frozen on the web. You can download and edit the project files themselves. Since you can download the result and open it in Godot or Unity (Unreal is embedded web atm), Seele03 can give you a quality core and a way for you to take things even further.

So how did it turn out? I've tried a lot of game generators, and Seele03's all-in-one tool approach and engine targeting stand out for me. Most others only generate three.js for web, or are more world models based on generative rendering. This is the rare one I've seen that is making games directly in game engines.

They've asked me to share the full process video with you so you can judge for yourself. As I watched the team work, I sat back and wondered at how it's all changing so fast: you simply... chat... interacting with a UI that is purely text, typing in wishes and directions, and everything else unfolds around you. I think they got pretty far in their goal gameplay-wise, which is where they really focused. The key thing is that after you have the final result, you also have the files to download, and you can replace anything you want to take it even further: animations, level graphics, basically anything.

If you want to make games for real engines on PC and console, then give Seele a shot at https://t.co/jJ9kvt0Ard. It's faster than vibing a game in Claude because of their game-design-driven model, their built-in asset generators, and the fact that you are not burning tokens re-creating an entire engine before you even get to your game. And every day they get closer to AAA... it's just a matter of time.

@SEELE_AI1125

[Future Stacked @FutureStacked]

@Grummz A playable wuxia action demo in five days is an intriguing starting point.

What would convince you that an AI-generated prototype is ready to become a real game project?

[Future Stacked @FutureStacked]

The iPhone Duo becomes a holographic projector in this demo built with Claude Fable 5.2.

[Future Stacked @FutureStacked]

This cinema app lets you preview the screen from any seat before you pay for it.

[The AI Colony @TheAIColony]

People who got rich in 2017 were early to:

• Bitcoin
• Dropshipping
• Amazon FBA

People who could get rich in 2027 are getting early to:

[AI Highlight @AIHighlight]

AI is getting dangerously close to Disney level animation.

[Future Stacked @FutureStacked]

StreamEval resources:

- Leaderboard: https://t.co/YxK4gtddFB
- Research Paper: https://t.co/Lf4kq3qeNF
- GitHub Code: https://t.co/Da5uIuBblT
- Dataset (Hugging Face): https://t.co/76Ywi2QRqa
- Dataset (ModelScope): https://t.co/gwJdcE45H3

[Future Stacked @FutureStacked]

Interactive AI video is evolving, but how do we compare which models handle real-time interaction most effectively?

That’s where StreamEval comes in.

Traditional video benchmarks score short, finished clips. StreamEval instead evaluates how well models handle real-time interactions, follow user instructions, and maintain consistency throughout a session.

The model must respond within the specified timing requirements without losing previously established context or state. 

The leaderboard ranks models across three categories:
• Avatar: digital human dialogue and avatar performance
• World Generation: scenes that respond to new controls as they play
• Video Editing: instruction and reference edits on a live source video
What makes it different?
• 5 interactive video tasks
• 300 scripted test cases, from 15 seconds to 5 minutes
• 10 task-specific metrics, scored from 0–100, covering task execution, state consistency, visual fidelity and audio-video sync

One interesting takeaway from the first round of 38 models: impressive visuals don’t necessarily mean a model can reliably follow instructions or maintain a consistent state.

Built around open-source, reproducible evaluation, StreamEval publishes its code on GitHub and its benchmark data on Hugging Face and ModelScope, along with the protocols and scoring rules.

Researchers and developers can inspect the implementation, review the scoring methodology, and reproduce evaluations when the necessary models and APIs are accessible. 

The research team includes contributors affiliated with Tsinghua University, NVIDIA, CMU, UC Berkeley, Princeton, NUS, and Reactor. bringing academic and industry research together around a shared evaluation framework.

AI companies and model developers can submit their API endpoints to be evaluated under the same StreamEval framework. It is a simple way to verify performance, spot weak areas, and see how a model compares with other systems under a consistent protocol for the relevant tasks.

The team plans to evaluate new and updated models and publish leaderboard results monthly. 

For founders and investors following the space, it offers an evidence-based view of how reported models perform on specific real-time interaction capabilities. 

As interactive video models continue to advance, reliable interaction may be just as important as visual quality.

Explore the benchmark: https://t.co/mUpozV30JF

What do you think is the biggest bottleneck for interactive video AI today?
#StreamEval #StreamEvalBenchmark #InteractiveVideo #AI