Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Jules: An asynchronous coding agent

jules.google

Post reply on HN
Print view View original

211–220 of 245 posts

Re: Jules: An asynchronous coding agent

#211

Post by bithive123 »

netdevphoenix wrote: post #174
xpe wrote: post #162

Earlier quoted context omitted.

Blame the system, not the actors. See a recent HN submission, The Evolution of Trust by Nicky Case: https://ncase.me/trust/ If there's one big takeaway from all of game theory, it's this: What the game is, defines what the players do. Our problem today isn't just that people are losing trust, it's that our environment acts against the evolution of trust. That may seem cynical or naive -- that we're "merely" products…

> we have to define what kind of society we want, predict likely responses, and build systems to manage them. Nailed it. At the end of the day, companies are automatons. It is up to use to update the reward and punishment functions to get the behaviour we desire. Behaviourism 101

You talk as though a company exists in its own right independent of the humans. This is a fictional way of thinking. This attitude of "if you want me to stop acting poorly, make me" is an abdication of all responsibility.

It's the idea that individuals and institutions must somehow fix society from the top down or the outside in, which history has shown doesn't work. No one is going to come along and make you be sensitive or intelligent, either you see the predicament we're all in and act, or you rationalize your selfish actions and make them someone else's problem.

Re: Jules: An asynchronous coding agent

#213

Post by achierius »

xpe wrote: post #162
bendigedig wrote: post #152

Earlier quoted context omitted.

This is the kind of attitude that leads to revolutions.

Blame the system, not the actors. See a recent HN submission, The Evolution of Trust by Nicky Case: https://ncase.me/trust/ If there's one big takeaway from all of game theory, it's this: What the game is, defines what the players do. Our problem today isn't just that people are losing trust, it's that our environment acts against the evolution of trust. That may seem cynical or naive -- that we're "merely" products…

Was he blaming anyone? He just pointed out the mirror of what you did: as the owning class acts one way, it will naturally produce material conditions that incentivize the working class to act in a way that would lead to the destruction/dispossession of the existing owning class (i.e. a revolution).

Re: Jules: An asynchronous coding agent

#214

Post by achierius »

rvz wrote: post #97

Notice how no-one (up until now) mentioned "Devin" or compared it to any other AI agent? It appears that AI moves so quickly that it was completely forgotten or little to no-one wanted to pay for its original prices. Here's the timeline: 1. Devin was $200 - $500. 2. Then Lovable, Bolt, Github Copilot and Replit reduced their AI Agent prices to $20 - $40 3. Devin was then reduced to $20. 4. Then Cursor and Windsurf AI…

Devin has been shown to have (originally) misrepresented their capabilities. Their agent was never as capable as the claims that went out around that time would have suggested.

Re: Jules: An asynchronous coding agent

#215

Post by myaccountonhn »

_pdp_ wrote: post #9

The copy though: "Spend your time doing what you want to do!" followed by images of play video games (I presume), ride a bicycle, read a book, and play table tennis. I am cool with all of that but it feels like they're suggesting that coding is a chore to be avoided, rather than a creative and enjoyable activity.

I think the copy is more for the authors themselves, since this is probably what they believe in.

"We're not replacing jobs, we're freeing up people's time so they can focus on more important tasks!"

Maybe helps them sleep at night and feel their work is important.

Re: Jules: An asynchronous coding agent

#216

Post by myko »

nsteel wrote: post #203
nine_k wrote: post #175

Earlier quoted context omitted.

Why, they'd happily delegate their own job if they've got to keep the proceeds.

Can you think of an example in history where labour was replaced with tech and the displaced workers kept their income stream? If a machine can do your job, (eventually) I'll be cheaper to use that machine instead of you and you'll no longer have a job. Is that not a given? Anyway, it was probably just a joke... so not sure we need to unravel it all.

I didn't get the impression it was meant as a joke:

"Every great venture capitalist in the last 70 years has missed most of the great companies of his generation... if it was a science, you could eventually dial it in and have somebody who gets 8 out of 10 [right]," the investor reasoned. "There's an intangibility to it, there's a taste aspect, the human relationship aspect, the psychology — by the way a lot of it is psychological analysis," he added.

"So like, it's possible that that is quite literally timeless," Andreessen posited. "And when the AIs are doing everything else, like, that may be one of the last remaining fields that people are still doing."

https://futurism.com/venture-capitalist-andreessen-jobs

Re: Jules: An asynchronous coding agent

#217

Post by d4rkp4ttern »

jacob019 wrote: post #202
d4rkp4ttern wrote: post #199

Earlier quoted context omitted.

Langroid has this kind of design (I’m the lead dev): https://github.com/langroid/langroid Quick tour: https://langroid.github.io/langroid/tutorials/langroid-tour/

Looks great, MCP, supports multiple vector stores, and nice docs! How do you handle to subtle differences in tool call APIs?

Thanks!

Langroid enables tool-calling with practically any LLM via prompts: the dev just defines tools using a Pydantic-derived `ToolMessage` class, which can define a tool-handler, and additional instructions etc; The tool definition gets transpiled into appropriate system message instructions. The handler is inserted as a method into the Agent, which is fine for stateless tools. Or the agent can define its own handler for the tool in case tool handling needs agent state. In the agent response loop our code detects whether the LLM generated a tool, so that the agent's handler can handle it. See ToolMessage docs: https://langroid.github.io/langroid/quick-start/chat-agent-t...

In other words we don't have to rely on any specific LLM API's "native" tool-calling, though we do support OpenAI's tools and (the older, deprecated) functions, and a config option allows leveraging that. We also support grammar constrained tools/structured outputs where available, e.g. in vLLM or llama.cpp: https://langroid.github.io/langroid/quick-start/chat-agent-t...

Re: Jules: An asynchronous coding agent

#218

Post by ddalex »

CraigJPerry wrote: post #130
CobrastanJorji wrote: post #69

So, you can assign github issues to this thing, and it can handle them, merge the results in, and mark the bug as fixed? I kind of wonder what would happen if you added a "lead dev" AI that wrote up bugs, assigned them out, and "reviewed" the work. Then you'd add a "boss" AI that made new feature demands of the lead dev AI. Maybe the boss AI could run the program and inspect the experience in some way so it could dem…

we're about to find out. This is our collective current trajectory. I am pretty convinced that a useful skill set for the next few years is being capable at managing[2] these AI tools in their various guises. [2] - like literally leading your AI's, performance evaluating them, the whole shebang - just being good at making AI work toward business outcomes

Just like a managers job

Re: Jules: An asynchronous coding agent

#219

Post by jacob019 »

d4rkp4ttern wrote: post #217
jacob019 wrote: post #202

Earlier quoted context omitted.

Looks great, MCP, supports multiple vector stores, and nice docs! How do you handle to subtle differences in tool call APIs?

Thanks! Langroid enables tool-calling with practically any LLM via prompts: the dev just defines tools using a Pydantic-derived `ToolMessage` class, which can define a tool-handler, and additional instructions etc; The tool definition gets transpiled into appropriate system message instructions. The handler is inserted as a method into the Agent, which is fine for stateless tools. Or the agent can define its own hand…

Love it, I did something very similar, deriving a pydantic model from the function signature. Simpler without the native tool call API, even though occasional retries are required when the response fails to validate. Will have to give Langroid a try.

Re: Jules: An asynchronous coding agent

#220

Post by babyshake »

nsteel wrote: post #140
oytis wrote: post #137

Earlier quoted context omitted.

Why not? VCs manage investors' money, not their own. If investors think AI is so great, they will have no problem delegating this job to AI, right?

I think it was a joke, VCs are happy to replace all jobs except their own.

VCs absolutely want to replace their job. Except for the part where they get paid. The actual work part they are happy to outsource.
Post reply on HN