Imo they explain pretty well what they are trying to achieve with SIMA and Genie in the Google Deepmind Podcast[1].
They see it as the way to get to AGI by letting AI agents learn for themselves in simulated worlds. Kind of like how they let AlphaGo train for Go in an enormous amount of simulated games.
That makes even less sense, because an AI agent cannot learn effectively from a hallucinated world without internal consistency guarantees. It's an even stronger case for leveraging standard game engines instead.
If that's the goal, the technology for how these agents "learn" would be the most interesting one, even more than the demos in the link.
LLMs can barely remember the coding style I keep asking it to stick to despite numerous prompts, stuffing that guideline into my (whatever is the newest flavour of product-specific markdown file). They keep expanding the context window to work around that problem.
If they have something for long-term learning and growth that can help AI agents, they should be leveraging it for competitive advantage.
Love the simplicity of your to to-do list syntax highlighter in comparison to todo.txt. That's more how my brain works, as simple as possible. Especially your take on the due date vs. date when you plan to do it.
Will definitely try it out.
I just tested the German – Standard layout on a Mac, it's the first one you indicate. Apple calls right-⌥ the Compose key, not AltGr, but works the same way, other than being transposed with the Command key from the familiar Windows style of layout.
There's also a keyboard layout just called German, maybe that one works the way you refer to? But surely it's not too much to ask that someone select the keyboard style they're accustomed to using?
3. Configure Ollama server to make sure it allows connection from grafychat.
That's not very helpful. Something along the line
Set the environment variable OLLAMA_ORIGINS to "https://www.grafychat.com" and rerun "ollama serve". Use your custom host if your using the self-host option.
"This study explores a violation of Newton's third law in motile active agents, by considering non-reciprocal mechanical interactions known as odd elasticity. By extending the description of odd elasticity to a nonlinear regime, we present a general framework for the swimming dynamics of active elastic materials in low-Reynolds-number fluids, such as wavelike patterns observed in eukaryotic cilia and flagella."
Thanks!! Great point; for now we're relying on S3+dynamo which many people prefer anyways; but state management is on the roadmap, we'll get to it soon
Those people have not experienced the :heart_eyes_cat: of GitLab's TF state store, which I find just a bazillion times superior to creating TWO separate AWS resources only for storing a bunch of JSON to make TF work: https://docs.gitlab.com/ee/user/infrastructure/iac/terraform...
> Drive. Unless you plan on pushing the car there
reply