_pdp_

I think this might work.

When I started working no the trading agent I mentioned above I wanted to see if it can be just a better investor over the long run. The intention was not to do high-frequency trading. As you can see most of the days it is not taking any actions. The losses where down to mistakenly setting the stop losses too close to the top. If it wasn't so careful it might have made some money tbf.

My gut feeling is that AI agents will be able to manage a long-term portfolio much better than a human. Though it is just a gut feeling.

If this worked it wouldn't have been open source?

Anyway, I have been running my own trading experiment and so far it has lost a bit of money. That being said I have not tried to optimise anything - just let it do whatever it wants. The losses are small and it might be able to recover later this year. Who knows.

The agent writes a blog about its progress here https://trades.chatbotkit.space/

I am thinking to output all the chat logs to HF as well for research.

You can run your own trading agents that communicate over a message buss in your own terms by downloading the CBK platform and running it locally with your own models. I have also shared my trading blueprint if you want to give it a go. https://chatbotkit.com/hub/blueprints/trader

GitHub. Then you symlink them into the relevant folders. You can write a skill to manage them for you across all harnesses as well.

Great initiative. Also add

https://github.com/chatbotkit/platform

...if it fit the project direction.

It is basically a sovereign AI platform for agent orchestration and automation and app building.

I like the looks of Aisle and what they stand for...

That being said you cannot compare a model with a specialised harness. These are two completely different things.

Am I missing something?

I get that. However, given that sessions are local and personal why tag it into the public source tree when it can keep a catalogue within the local session metadata with the exact branch, tags and refs. Sounds like easier less intrusive thing to do.

Respect

Because Linux is not an OS.

I really hate this attribution nonsense ...

On one hand I guess it is fine to automatically tag that the commit was AI-generated. On the other hand, this will happen even if you audit every line and add your own changes too, which means the attribution is misleading.

My point is that it is a great way to learn and be more appreciative of what is done and why - network stacks, graphic stacks, web, etc - before being too critical of them.

Besides, statements such as "$7B for a proxy" or "it is just a wrapper" or "we can build it in-house" mostly come from those that have not had any measure of the scope due to not having any real experience with the said problem.

Everyone should build their own ** is the general advise I would give just about anything. Otherwise it is hard to appreciate how much effort it goes into things that we take for granted.

Everything looks simple from the outset until you try to build it yourself.

Most security bugs are due to improper string validation and use.

Nothing to see or learn from... move on.

Lol. Is anyone actually taking this seriously?

We are talking about the process but not the result. I think the results will speak for themselves.

IDK In my experience humans do go into doom loops.

Well if the recipe requires access to some secret ingredient it may as well resort to hacking to obtain it. ;)

I might have a different definition of "rogue" but to me it means when you go outside of the rules/norms ... and this is happening all the time.

All computer primitives are relatively straightforward in pure form and vastly more complicated in real-world scenarios.

> we are just a couple of steps from the possibility of a true rogue AI

No no. We are not a couple of steps away. This is happening. AI is already used for hacking and creating a harness that makes this fully autonomous is relatively straightforward.

detect, slice & dice, re-prompt

Doom loops are as much of a model problem as it is deficiency of the harness. I have not seen any other open source harness that deals with them except the one I started because of this obvious gap.

See my other comment with examples where 0x Alpha is working non-stop on various projects with zero problems.

Ox Alpha has been running on auto-pilot for the past 5 days on various experiments.

Very impressive model.

Here are some examples, open-source documented and the data available in HF datasets:

https://openzot.github.io/whetstone/ - https://github.com/openzot/whetstone

https://openzot.github.io/arcade/ - https://github.com/openzot/arcade

https://openzot.github.io/machinery/ - https://github.com/openzot/machinery

I've seen a lot of comments here and elsewhere about why is a proxy worth $8b. It is a bit more than a proxy.

It is a known brand for model routing and it has developed means to route model on their own and 3rd-party infrastructure at scale and without much issues.

Why Stripe? Who knows. I don't see much synergy but why not? If anything it is a way for Stripe to get into the market with agentic payments which may come to be worth a lot more than $7B.

It may come across harsh but IMHO the only interesting thing about this project is that it is written in Zig. That's it.

Everything else in the harness is largely the same just Vercel-flavoured.

The portability benefit is also a bit over-sold imho. I wrote a harness in Go and it is as portable as this ... in fact it deploys straight into Vercel's own sandbox environment on demand without any issues.

That being said, did you say GLM 5.2 free? I need to look into that. GLM is remarkably capable model and that alone is worth using the harness in my books.

Considering that ripgrep, git, and, you know, other dev tools are part of the toolbox, then why ship them inside this executable? And, furthermore, if you ship them, then why stop there?

Plenty of previous art.

Very interesting. Thanks for sharing.

If you can spin up a new software project with little to no effort, the most important part of the job becomes making the right decisions and this is where we will be spending 99.9% of your time.

Full disclosure I'm a founder of cbk.ai and chatbotkit.com, where we've built reusable agent components. I have never really saw flowise, n8n the now defunct OpenAI agent builder and other similar products as competitors and I wanted to describe why.

I think workflow builders are mostly the wrong mental model for AI. A graph of predefined nodes and edges is useful when you already know the exact sequence of work.... but the point of an AI agent is that it can decide what to do next from the current context.

IMHO reusable building blocks still matter enormously. You should not have to recreate an integration, a capability, or a piece of reliable deterministic logic every time. It is wasteful and you cannot oneshot it. Even with coding agents which we use all the time system to system connectivity is more complex than writing the code. But those should be available to the agent as tools, mechanics, and features that are not wired together as a visual workflow of sorts.

A product can look graph-like at a glance though, while avoiding making a workflow builder. I think this still matters in order to put a mental model around it. But the mechanism should not be graph / workflow like in practice.