soundworlds

There is an enormous amount of time that goes into ticket and other triaging admin across industries. Its use cases go beyond that, but that alone would save a huge number of person-hours.

I love that, within days of Jev being released, we have a flurry of open alternatives. This is kinda the open-source dream :)

I noticed my YouTube recommended searches seeming to change based on what Claude has just recommended me the other day. I do wonder how much of this data is being shared between the companies.

I this is part of a larger issue of our economics rewarding immediate results, but not good process.

Basically, as long as this is the mechanism by which people earn money to stay alive, the world will always be optimized towards results.

For me the tell for AI-generated-anything is too much detail

Like, it just throws everything in there, because it doesn't have the same friction that humans do.

Blocked the a few attempts I tried, usually by changing the amount of fencing:

- Let the player jump over walls

- Let the player teleport the flag to them

- Make the world 3D

Interesting, I shall have to try this on other software!

Yeah I was skeptical about AI until I learned (through hands on testing) that an AI can fail 9 times and get it right on the 10th time - in like 1/10th of the time it takes a human to get it right. It's all in the iteration.

Turns out geopolitical coordination around AI already exists. It's just not American

Related: I've started getting my agents to create and use a KANBAN.md as a source of truth for software projects. It's turning out excellent for agent handover, and tracking design decisions all in a single place.

Yeah fair point. I believe he truly has humanity's best interests at heart.. And yet he keeps speaking as though everyone still sees the US as the shining beacon of the world.

That's simply not true anymore, and his messaging will continue to be undermined until he treats other countries as equals (whether that's China, Australia, or anywhere else).

> They have shown time and time again that they are not to be trusted, yet, the main ask is to trust us, only us.

Exactly. For everyone outside the US, we see a tech industry that has spent 25 years moving fast and breaking things AND elevated Trump to be POTUS - which has destabilized the rest of the world. This is after 70 years of the US invading and destroying small countries, often without plans or robust reason, all in the name of "democracy" (ask any other country if they feel like they had a vote in the US's actions)

Why the hell would we put our trust in a few AI Labs in the US, when this is the legacy they will be reinforcing? It has to be Open models - for the people. No more of this US-paternalistic bullshit.

I was involved in this research a few years ago: https://metamiditoolkit.com/research-paper/

One of the things that struck me was that most people have an area of their workflow that they would be happy to hand off to an AI, so that they can focus on the thing they love. However, that area is different for each person. One person's grind-work is another person's love-work.

This is why it's so hard to build consensus around where the "red lines" are in this space.

Funny, I'm now finding my first thought with new hardware like this to be "how good is it at running AI models locally?"

Yeap, those are the games that inspire me most!

A large chunk of the population don't use Google to search anymore. They go straight for the GenAIs. That's got to be scaring a lot of managerial folks at Google.

100% - the work came from the people, it should go back into the hands of the people.

I also think if Anthropic and OpenAI had been releasing Open models along the way, people wouldn't be nearly as suspicious of them.

Check out DeepSeek v4.1 Flash, just released, and I am enjoying it more than Claude. No one can run it locally.. Yet.. But give it a few years and I wouldn't be surprised if models of that capability were standard for programming laptops.

While I respect Dario taking responsibility here, this line disturbs me: "The US and other democratic governments attempt to coordinate with authoritarian governments"

We all know he is talking about China, and I'm pretty sure China doesn't appreciate being called "authoritarian". I am sure he doesn't mean it, but in all of his essays, his language around non-US nations always disturbs me a little bit..

--

Also Dario, if you happen to read this, I want you to know that I have loved using Claude Code for programming. But I am now using DeepSeek v4.1 Flash - simply because it is the same good experience, but Open Weights. Making the Open Model space succeed is where I am investing my time - it's giving back to the people, true and simple.

Even if OpenAI didn't use their training data, they heard about one of their customers working on the problem of their career, and then undermined them.

Does OpenAI just see this as fair game?

Regardless, they started working on this problem after hearing that one of their customers was already working on it. It almost doesn't matter about the training data. This is the provider you are paying undermining your career.

We all just need that one friend in our group who will pay for it. Then the rest of us can play with it briefly until the novelty wears off.

Yeah I'm pretty sure "load-bearing" is their watermark

It sounds like a friend who's just learned a new word and wants to use it in every sentence

Godot has that potential, and can export to web. Though I do suspect LLMs have just come and taken that spot.

Totally! The web has felt like it has lost a lot of heart with the disappearance of Flash. The key part of Flash is it let people aren't steeped in web design make "technically bad" interactive web applications, which is one reason you got so much variety.

The great news is that LLMs actually open this box back up. People with basic understanding of web can iteratively generate single-file HTML pages, AND LLMs are good for including accessibility guidelines, generating alt text, AND the source code is completely viewable.

In theory, this can actually usher in an even more creative time that the Flash days.

Can confirm, I had never heard of Navier-Stokes before this fiasco. And while I suspect OpenAI decided it was worth the risk for the public display of capability, this proves they are now directly competing against their own customers.

The future I'm after involves Open models controlled by the people, not a few SC venture capitalists

I think this is where people will have to let go of the ego of being the "sole author" of a solution for us to move into the next era of human flourishing.

I think you may be underestimating how crucial a good User Interface is for widespread adoption of a product..

Reporting back - project files are still huge. I suppose that is just a side effect of it bundling the audio directly into the project file.

Just makes the files very hard to share with others online though..

Hahah - I woke up this morning going "I want to use something like Audacity, but doesn't create such enormous project files".

Hoping they've fixed that, looking forward to it! UI looks great too!