tizkovatereza

Hi all, I paid ~$50,000 for a booth at conference for the first time and had too many not so obvious things I wish I knew before that.

Hope I can learn tips from others too. What is your experience with conferences, especially in tech?

Hope this is helpful for others and happy to discuss

t

I agree with this, the logs are important for our users already, especially for the ones on big scale.

I definitely agree, but would say this is more "weak recursion" than strong one because you still need human intervention for that self-improvement, no?

AI is a rare example of self-improving (recursive) technology. What are other examples similar to that in principle? Do they even exist?

There is a new AI agent that generates Next.js apps based on prompt.

Features: - Plans frontend component architecture - Generates code in real-time based on commands - Deploys code to a secure sandboxed environment for testing - Works with any LLM - It's all built on top of Cerebrium and E2B

Complete guide to building the agent: https://www.cerebrium.ai/blog/building-a-real-time-coding-as...

Source Code: https://github.com/CerebriumAI/examples/tree/master/8-applic...

WebDev Arena lets you compare models like Claude Sonnet, GPT-4o, Qwen, or Gemini for free.

It compares the models on web dev tasks of your chice (e.g. "create a productivity tracker app" or "make a data analyst dashboard"), lets you vote for the best result, and then reveals which model created which result.

It's built with E2B as the runtime for generating the apps by executing the LLM's code in the sandbox environment.

We just launched a secure virtual desktop made specifically for LLM-powered computer use. The difference from Claude computer use is that: - It works with any LLM

- It has high level of security, because each sandbox instance is isolated from the others, and it's isolated from user's os.

The E2B sandbox in general can be customized with any dependencies or pre-installed packages.

Some features:

- Python or JS/TS version

- Allows streaming of the desktop screen

- Supports mouse & keyboard control, opening files, screenshots, running bash and PyAutoGUI commands & more.

Here is the Desktop Sandbox repo: https://github.com/e2b-dev/desktop We also added an example of computer use, using the E2B Desktop Sandbox: https://github.com/e2b-dev/secure-computer-use

I will be happy to hear any feedback if you want to check it out.

Thank you!

Hey @janoc, thanks for this point.

This is just very simple example project, not meant as a full product to offer people or bigger scale.

Of course if the usage happens to increase, we will update the program to prevent the problem you are mentioning. (E.g. limit to private repo, limit the number of PRs etc.)

I am very new to coding and wasn't really aware that people are creating PRs just for credit in that extent.

Hello, This is just very simple example project, not meant as a full product to offer people. As far as we are monitoring it, users are testing it on their own repositories. Of course if the usage happens to increase, we will update the program to prevent the problem you are mentioning.

Hey huksley, not at this moment, but thanks for this idea! I worked on this project and will implement this in the future, shouldn't be difficult.

I would say in the end you will have to build a whole product around your agent, no matter what. So the same for GPTs or other agents to make them reliable on prod.

AutoGen is very popular recently among frameworks for building agents.

In their paper, they compare with other frameworks like CAMEL or BabyAGI, and one difference that stands out is execution-capability.

I wrote something about its potential limitations, plus added a quick guide for code execution.

Happy to discuss!