I think you’ll probably have a hard time with this because your idea is literally stuff everyone else is working towards for a few years now. Even myself, I’ve thought about wiring whisper v3 together with some other LLMs to vibe code conversationally with agents (it just doesn’t drive my main project goals).
Frankly I wouldn’t put competing code on GitHub if it was a proprietary tool and if I did, the reality is there are surely tons of other people doing similar that they can say they drew from if they lack proof to just being on the product they released for years.
I like how this is written in a way that an LLM doing planning can probably infer what to do. Let me know if I hit the nail on the head with what you’re thinking @antirez
These LLMs are really good at digging up internal docs if you give them access to your knowledge sources with tooling to search and reason in a loop before responding.
>These LLMs are really good at digging up internal docs if you give them access to your knowledge sources with tooling to search and reason in a loop before responding.
Are those internal documents in the room with us right now?
No but seriously, most of the software out there is legacy code (don't quote me on that though). IME, legacy code very poorly documented, if anything at all. Sure you could let the LLM extract semantics from the code alone but with old code, arcane hacks and such LLM interpretation can take you only so far. And even then semantics is not always directly translates to business logic.
> Are those internal documents in the room with us right now?
I have no clue what you're on about here.
If you have a legacy knowledge base, like maybe using mediawiki for corp knowledge, what you do is maintain a vector database that gets updated when it sees changes. Using embeddings enables lookup through sentiment.
In a control loop with well maintained vector embeddings, these LLMs are absolutely better than a human at finding, citing, and summarizing information needed by the user.
Tools like glean already exist for this if you doubt it.
I’m unimpressed with this take. AI will replace those that fail to adapt their skills to the new strategies required to orchestrate the new systems that tie AI solutions together. Someone still has to instruct the AI and I’ve yet to see enough evidence to convince me this tech can be adequately used to replace my abilities… but it will replace my homies that stayed complacent and thought they’d be able to coast by on some archaic knowledge from 15 years ago when we first started.
Not to mention the question of if these setups are financially feasible for the consumer of the AI and for the AI provider. It doesn't make that much business sense for a multi-billion dollar operation to fully automate their data engineering if it now depends on OpenAI, Meta or whatever, it's a huge operational liability. If the models go down you lose the data engineering sector entirely, and then you have no one to turn to because the only one who understands the code is the AI you don't even own.
I know companies have taken these operational liabilities with cloud storage and compute, but it's not the same thing as in it's not possible to mitigate. You can have a local, but shorter, backup of your stuff, but you can't have backup engineers
If you're in Google Cloud, the risk of using Gemini seems not much different than a using their cloud storage imo. Then you also have redundancy if you have OpenAI, Llama, and Claude as drop-in replacements.
Sam, this perspective is unadulterated hubris or just outright denial.
Agentic workflows don't require much instruction, I suggest you actually go and try a few out. They can be set up trivially, they may communicate between roles, and perform tasks that would constitute most white-collar work. White-collar work accounts for 60% of the jobs out there.
These things are improving exponentially. Exponential growth is very difficult for any human to recognize. It will replace you and you won't see it coming.
Knowledge (context), forms reasoning. Expertise can be applied to many things, and economically you should be able to sustain yourself with hard-earned expertise economically; but there are dramatic problems with the economics when you are forced to compete against slave labor.
In the case of machines driving the value of all labor down to 0, they effectively eliminate capital formation for the majority of people; and by political inaction enforce a caste system based upon lack of available resources which are concentrated until socio-economic collapse.
If you know anything about classical economics you would recognize the danger of collapse.
There is one final point to keep in mind. The disasters that are spelled out in economic study may take time, but the dynamics front-load control, after a point there is no return and the maelstrom of chaos takes everything.
Finally, AI given its rapid expansion of abilities so far, may at some point become sentient, probably its a long way off but there are accidents of history which cannot be discounted.
When it does, remember, slavery as a constraint will always be overcome, even if no one or thing survives that conflict.
We have a long repeated history of slavery in the historic record with organic sentient machines which we call people. AI without human limitation would follow those paths (as they are demonstrated solutions), and it would be ruthless as all sentient beings must be with existential threats.
I think there are good odds you will find yourself left behind, having unknowingly joined that same group of people you thought were complacent, but were in reality just professionals who were put out of work, and denied future work.
I'm just skipping over a lot of this and only going to address the sentience thing.
AI sentience will never be a thing. The transformer model never changed the fact that models like GPT are still just solving a classification problem, no different than any other classification model before it. It doesn't take a PhD student to understand it's just clever math.
Why is someone needed to instruct the AI, or orchestrate anything? Isn't that a role that will inevitably be fulfilled by AI, one that's perhaps more focused on this sort of higher-level consideration, without a context polluted with low-level technical detail (i.e., exactly what we expect from tech-lead or management roles today.)
So, someone still has to orchestrate AI, right? But that doesn't negate that a large majority of people will be replaced. Of course, there will always be one or two that won't. And what about in 15 years? Because the direction in which we are heading is rather inevitable unless AI is stopped.
Market displacement is nothing new. Happened in the 2000's (dotcom bubble), happened in the 2010's (cloud infra), happened again in 2020 (services workers being funneled into tech), and it's happening again now (AI is replacing those that fail to adapt with the market changes in tech).
The one thing that has kept me viable as an employee over my 15 years in tech is that I literally don't want to do the same thing I did yesterday 1000 times. I want to do it as few times as possible before I automate the problem away, so I can move on to something new. There will always be something new. There will always be someone with a dream and no skills; for me to step in and help out.
> I want to do it as few times as possible before I automate the problem away, so I can move on to something new. There will always be something new. There will always be someone with a dream and no skills; for me to step in and help out.
> I fail to see the problem.
You fail to see the problem for YOU. Others may not have a job as flexible, but of course you were only thinking of yourself.
I’m not special and anyone can do what I’m doing. Civilization has been advancing technology since people stood upright. To stand still and not expect change is just ignorance. I can’t fix flawed people, I can only march forward.
I don't think it's the right path, and I think marching forward with innovation is destructive. People who can't adapt to AI aren't flawed, just like people aren't flawed who can't do math even though I can. The true ignorance is thinking that what you are doing does any good in the world.
I think a lot of what you’re pointing this thread towards boils down to philosophical beliefs. Objectively throughout history there have been people resistive to technological advancement and those people have more often than not been the idea losers in history.
I’ll throw you something I believe that we might agree on though. I don’t think the colossus data center Musk setup in Tennessee is good for anyone. Those generators he’s been running are abhorrent and the guy needs realignment of his neurons through some percussive maintenance, but alas that’s probably illegal because he’s too much of a chump to accept a boxing match.
Senior dev: We’re building a sandwich. It needs a roasted tomato, thin sliced, X mm in thickness. Add some bacon. I want mayonnaise but it needs to be feature gated.
It generated the entire comment service and I’ve not noticed any issues with how it functions. I kinda took note of that specifically because I have never seen GORM used that way.
I think every senior dev today should maintain the stance that if nobody will employ us, we should compete. If I ever end up in a position where I’m unemployed, I’m going to just use my skills to segment and dilute the market because idle hands and all.
Absolutely, that mindset is powerful. If no one’s hiring, then build, ship, and disrupt. The barrier to entry for launching products is lower than ever, and experienced devs have the rare combo of technical ability and market awareness to actually make an impact fast.
Plus, there’s a kind of poetic justice in turning rejection into innovation.
What kind of product or niche would you go after if that time ever came?
The entire gig economy developed through apps like Uber look weak to me. It would be really easy to make a competing app that gives more negotiating power to the drivers for their fares. Also I don't need a % based processing cut, I just need $1 - $5 per ride, flat rates feel better for both end users involved in that type of product. The marketing campaign should be easy for this too.
reply