Zeruxe

Thats what I was thinking, this is not a breakthrough?

This is quite interesting.

5k is nothing for the amount of headache you just saved them..

Okay thank you maybe I should of taken a better look at the guidelines.

Yeah I did contact them a month ago - no reply. Then again 2 weeks ago - also no reply.

Thats why I made this post.

Okay so one bad habit I have developed because of agentic coding is blind shipping.

By blind shipping I mean a scenario like this: I want to implement a functionality but I describe it on an abstract level to my agent, my agent implements it in like 5-8 commits and pushes the solution - the solution works! im happy and go to the next problem.

Now yes I know what problem is solved but I am still not exactly aware of how it was solved, this happens far too often to all types of developers and engineers.

Now if you are a hobbyist and doing this for fun then you don't really care about the implementation you are just happy it works and continue on with your day.

But as an engineer you are responsible for the code you ship and sometimes you ship at speeds that makes it impossible to review the code. So I wanted to make something that was as simple as possible to use but also as effective as possible to review your code.

So I puzzled ownit together. Ownit is simple and straight forward. You add a functionality whether it's you or your agent. This gets commited in X amount of commits. When you soon come to push this. You get an interview, the interview is based off the git diff of the commits you added.

It will ask you things about the commit, why you chose this , what it solved, how it solved it. Simple questions that make sure you understand what you are doing.

There is also a badge on your repo showing how many commits you actually interviewed for and how many you skipped, to sort of show how authentic your code is and if you chose to interview for it or just blindly shipped it.

The way to set up Ownit is very simple and FREE with a NVIDIA nemotron api key you get at nvidia models. Ive had a few devs try it and honestly the feedback was great.

Please let me know what you think if you try it out and if it actually improves your understanding of what you ship.

Surveillance is probably one of my biggest concerns with AI. The ability to monitor millions of people isn't something new, but AI makes analyzing all information drastically faster and more efficient.

And what about the BIAS we mentioned earlier in the thread? Wouldn't that influence the surveillance? isn't that a bad thing? there is just so many things to consider.

I just find it hard to understand the tool arguement from time to time. This isn't like when spinning jenny came out, or when the calculator came out.

It's so much better,smarter,more effective and affects so many different parameters.

That's a way to put it, is that why they want to call it superintelligence now ? haha

Yeah but how would you realisticaly enforce this ? global regulation?

Let's pretend US and EU agreed to slow down development, what would stop China or other countries outside "west" from continuing?

I think that international coordination would be very difficult when AI is a national competition.

Let's say the big companies can sustain the losses for now, surely at some point they still need to justify the massive losses. If progress plateaus while the infrastructure compute costs just keeps going up. Wouldn't that technically eventually even force the biggest players in the game to reconsider?

Very well said..

But why would they stop when they have a clear path up? They don't care about the people they just care about growing their wallets. I don't see them ever pressing the stop button aslong as results are there, also I do think that results will probably deteriorate, as I would like to believe transformers will hit a limit or rather a bottleneck, perhaps several bottlenecks. Where progress will massively slow down.

Do you think companies are deliberately exaggerating AI's capabilities to attract investment?

I think the economic questions are easier to evaluate. If AI consistently delivers better results at lower costs, there's a pretty strong argument for adopting it. But that is not the case for the large frontier models that actually produce the "best" quality out of the rest, they are just too expensive.

But the smaller models are somewhat profitable as they require almost no comparable compute.

The surveillance concern is definitely justified and I think it is already happening before our own eyes..

Trust me, I sort of agree with you and we are in the same boat.

But that is also the question I ask myself.. Is it really just a tool ? can we really use the same comparison as when the calculator came out ?

I am not opposed to using it, I am rather concerned about where this road will eventaully lead us.

I'm not arguing that AI can't improve productivity or code quality. I'm questioning what happens when developers become so reliant on it that their ability to reason through problems independently starts deteriorating.

9% believed it does more good than harm.. haha

Just looking through that quickly - "56% of AI experts expect AI to have a positive impact on the US, compared with only 17% of the general public."

Are the "experts" being optimistic ? or is the general population being too concerned?

1. Agree and it sounds horrible to think about, I really don't like the counter arguement "it's creating new opportunities".

2. I don't think automating tasks necessarily means eliminating a profession, but what AI can do is unfortunately speaking more than just automation.

3. Do you really think so ? I mean what do you think about the AI companies losing billions and billions every year, don't you think at some point they will just not have it anymore?

totally! and this is what I think can be really dangerous if we try to implement such things trying to make it "conscious".

sorry but are we really gonna be calling it superintelligence now ? haha

I totally agree with the benefits of the smaller models, but it's the large frontier models, potentially or accidently removing the security barriers from those can cause things far worse than the "hugging face" incident..

1. To answer your first point, don't get me wrong again. I understand what you mean by catching human errors and review. But it also opens other doors to "ai problems". AI definitely introduces it's own errors, there are cons to using it not only pros.

2.But that's exactly my point, Engineers feel forced to adopt AI to remain competitive, but high adoption rates in engineering professions don't neccesarily mean positive attitue towards it. The companies just want shipping..shipping.. They don't even care if the engineers understand the concept of what they're building.

I feel like the use of these habits will be biting them in the butt a few years ahead.. people should really be worried about these new workflows more than what they currently are - do I make sense?

If this is true(which I hope it is) it would certainly be calming, not because im against it totally.

I just think turning the blind eye will hit us like a big truck in a few years.. It really is about time..

but then, who is to program these so called morals and ethics? I feel like there would be an unavoidable BIAS when integrating morals into it..

I mean I definitely see your point of view, I just don't think those things are VERY contradictory. You can find it useful while still being concerned about its long-term consequences.

I think people or rather especially Engineers are just afraid to fall back and not be able to catch up if they completely drop AI. Which is why they use it even though some of them might not be very willing to.

Im curious to know if people that fear it, also don't want to see the development go well, or if they are still for the development. Perhaps the former right?

That's fair. A lightweight self-hosted model can't be so harmful(yet).

Perhaps more about the huge frontier models being developed at openai or antrophic.

Or perhaps the controversial stuff being associated with Palantir.

People are starting to ask questions, I mean a lot are just worried, I had so many people tell me they were pro-ai a few years back, but now they just have that worried face whenever you mention it.

I mean you are probably right, I think I came by a survey that showed that the concern is higher in "wealthier countries"

What more interested in is if people are just against AI itself, or just worried about how it's being developed and deployed.

Besides, "the worried west" also probably takes up the majority of the AI usage.

I wanted a way to perform market analysis easier, I just didn't have a way to do so fast, well I tried using the Gpt and Claude chat interface, but then I have to keep track of everything it spits out, and I have to genuinely prompt in such a detailed manner everytime for it to yield good results, because prompting is a skill whether we like it or not, the LLM's can't just assume what we want or need and get it correctly everytime.

Anyway, I wanted to think of a way to sort of speed up the prompting phase but as I wasn't sure how I came by Jules, Now Jules is a google model that works off my requests rather than tokens. So it has 100 requests per day, and in each request it could produce thousands of lines of code or text.

So then I constructed a very detailed prompt and sent it into Jules with different topics I wanted to market research, could be literally anything you want to investigate. But one thing hit the ceiling, I had to keep changing the prompt around everytime so I still had not really "automated" or saved "a lot" of time.

Then I got an idea of getting another agent to instruct Jules for me, so I just brought in an API, now at the time and still, NVIDIA offers free gpu compute models, so I just threw in one of the models and wired it up to nemotron 120b(you could also use any other api), the orignal prompt is very detailed and man-made, but nemotron sort of changes it around depending on the subject to be explored.

So I ended up with an orchestrator, and now all I do is interact with my orchestrator what subjects I want to research, send it in and the LLM will parse it and delegate it to Jules for me, all I have to do is review the pull requests made in a different repo.

Now for those of you who have made it this far reading, this can be seen as "slop generation" although I would like to point out that it is only slop if you let it be, if you construct a good prompt you will get a good result, and the output doesn't have to be taken to the fullest, you could also take it as inspiration or a lead to what you actually want.

The point of this is just to speed up different types of research, whether it is a litterature reivew, market analysis, research paper,experiment, MvP, Proof of concept - you name it.

It is also open-source and available here: https://github.com/Yoosseph/jules-research-factory

Gotcha, ive reached out!

Sounds interesting,where can I reach out to you personally?