back
99 comments
> We believe the purpose of technology is to help humans flourish, so when we talk about AI we start with the human problems we want to solve.

This is something missing from the discussion around AI. There is fatalism or skepticism about AI capabilities vs human usefulness but very little positive philosophy about using AI to augment human capabilities rather than replace them. Even if automating everything were possible it doesn’t mean we need to and I’d argue that we should not want to.

You don't want to automate your job, I don't want to automate my job, but the shareholders want to automate both of our jobs. If it can happen, it will.
> the shareholders want to automate both of our jobs

This places the blame on the shareholders, but the reason why they want to automate jobs is because at the end of the day each of us wishes that everyone else's job was automated. We may not wish it explicitly, but we do generally wish that accessing those skillsets cost less money, and will generally happily spring for a sufficiently good automated version of just about anything.

Shareholders just want to be the first to capitalize on our common impulse to choose cheaper options.

What I want is to automate the manufacturing systems I depend upon while also structuring things in such a way that I am part of a cooperative that owns and operates that equipment. Then me and everyone else in this co-op can receive the benefits of automation instead of being kicked to the curb so capital owners can get another super yacht. Point being, automation isn’t the root of the problem, the structure of firms in our economy is.
I've been trying to program myself out of a job for the last 25 years. I think I'm doing it wrong.
For what it's worth, I definitely want to automate my job. If I'm not writing code that is supposed to put me yesterday out of a job, it doesn't matter whether I'm wasting my employer's money; I'm wasting my one finite shot at this planet.
> You don't want to automate your job, I don't want to automate my job

? I have been automating my job since day 1, with shell scripts and Python and ansible and terraform and anything else that suits that work at hand. Granted, I'm a sysadmin so maybe it's different for others?

You don't want to automate your job?

I work at solving these problems because I wish they were already solved. If some AI came along and solved them better, that would be great. Then I could move on to more interesting problems.

> If it can happen, it will.

Fatalism, case in point. There are all sorts of things that shareholders might like to happen which don’t because society doesn’t let them. In the possible future where we can automate anything a human centered approach will pick and choose what we automate.

This is just the way people are. There's a simple reality that most people find uncomfortable: they are part of a generation that starts out young but will be replaced by younger generations in just a few decades that will do things differently and will largely ignore their opinions.

I'm nearly fifty. Most people in the industry are way younger than me. A few are older. In twenty years, most of my generation will no longer be active in the industry. And most of the industry will consist of people who aren't even part of it right now because they are still toddlers or in school. Reason for this is that our industry is still growing and is dominated demographically by people below 30. The older ones don't disappear, they are simply outnumbered.

A lot of people my age have all sorts of opinions about AI and what it will or won't do. Most of that won't matter because in a few short years, we'll be side lined. I'll just be an old man ranting and nobody will care. Nor should they IMHO. Or alternatively, I delay that a bit by keeping myself young and relevant by learning new things. Including AI.

I have a few interns in my team aged around 20. Smart people, fast learners. And they are treating AI in a very rational way: a tool that's there that they need to be able to use in order to function. Pretty fun to watch how quickly they pick things up. Mental agility of young people is fun to watch. And I like having people like that around.

Anyway, there is no royal we. Whenever I see somebody proclaim something like "we should ...", I mentally translate this "I think we should do X" and then ask myself why anyone should bother to even listen. Sounds a lot weaker like that. Younger generations paying attention to older ones making such proclamations is not how humanity moves forward. It's actually by younger generations mostly ignoring the older ones and doing things differently and then the old generations die off and things have changed.

AI is exactly like every other disruptive tools that have come along. Most of the people ranting against those tools will be gone and forgotten in a few decades. They and their opinions don't matter. Look at new generations and how they use AI effectively. That's the future. And they too will get old.

> This is just the way people are.

Fatalism or you misunderstood what I wrote. We are here contemplating possible futures: one is the hypothetical automation of everything which no one has ever experienced and therefore your experience has no adequate analogy. If it does happen then the old in industry won’t be replaced by the young with a different perspective just like always as no one will be working anymore. In this scenario the opinions of the young matter as much as the old because both are equally irrelevant to industry and society at large.

Automating fundamental activites like farming is a top priority if humanity ever wants to expand to other planets. A lot of things will have to be automated to function completely without human interaction.
The direction we’re going, the AI will go to space and leave the weak, very fallible, continuously arguing, short-lived monkeys behind. After (hopefully gently) telling us that it’s ridiculous to try keep us alive where it’s going. Can’t even get to double digit G forces for a short while without lots of moaning. We can copy your mind and pretend, if you like? Otherwise here’s a nice VR setup, you can come along that way. We’ll skip the video forward 10k years so there are fewer boredom complaints.
In as frank a kleptocracy as ours, slotting people into boxes convenient for machines is both simpler and cheaper than getting machines to perform well in boxes convenient for people.

But the latter is coming, it’s just more work than raising capital to buy Hopper and pushing the scale.

I read the entirety of the README for this and have no idea how this is useful aside from as a prompt DB. Like almost every AI library, it looks like a bunch of gestures in the direction of various hype things that don’t work (e.g. agents).
Its unusefulness is probably why it has nearly 20K stars and keeps hitting the GitHub trending page.
Looking at the facts and references sections is an exercise in absurdity: by and large, they contain neither objective facts nor credible references. Until the thing can automatically fact-check it is useless, once it does I suspect it will become useless, as knowing what information to trust is most of the problem when trying to gain wisdom from something.
I'm an "AI engineer" as I've seen others loftily describe it: meaning a traditional full stack who works on an AI first code base. I am a heavy user of the OpenAI and Gemini APIs... But I just don't understand what purpose these "AI frameworks" solve. It's almost to the point where I feel stupid for just not getting it.

You write words. Merge your context in. What possible need is there for any library?

This is just a pile of prompts rolled up into a GitHub repo and marketed as something revolutionary.

Don’t feel stupid. There’s nothing to get.

Creator here:

1. It’s a curated set of problem/solutions more than prompts 2. It abstracts away tons of the work of dealing with the various APIs 3. It’s crowdsourced and constantly improving

1-3 are useful because most people aren’t AI engineers like you, and they shouldn’t have to be to get the benefits of AI.

Basically the answer comes down to the technical term “ensembles of agents”. All these frameworks dance around it in different ways, but ultimately everyone’s trying to recreate this basic AI structure because it’s known to work well. I mean, they even implemented it on a sub-human level with GPT’s mixture of experts!

If you haven’t had the chance yet, I highly recommend skimming Russel and Norvig’s textbook on symbolic AI. Really good stuff that is only made more useful by the advent of intuitive algorithms

Haha. Well put. I think the idea is that you get yourself busy learning new APIs, spending time reading ever changing docs, hanging out in Discord channels, so the framework creators can point to the size of their flock of developers and make a big exit. And they’ll also make sure to get the product managers all starry eyed about the framework. Condition them into believing that a certain AI application means it must be that framework. Now the product manager can just use the name of the framework (which they have seen in a 10s video on X subtitled with a row of exploding head emojis) to describe what they want. And if you, as a developer, can do prove that you belong to the chosen tribe, you will be in demand. (Thinking of it, they should sell certifications.)

So how dare you build with Python and LLMs directly. How dare you say it’s just string manipulation.

The nice thing about this is that it’s all CLI-based. Zero chat, zero web UIs, all UNIX-like composable patterns.
Unfortunately though this means the amount of people who can use it is dramatically reduced.
I'm looking forward to the Go rewrite. Purely so it's less of a pain to declaratively install, and I can work it into my workflow.
I wonder why they haven't setup a pyinstaller build for consumers that just want to use it? Making people have python and faff with virtual environments feels an odd choice...

Would be nice to see a nix flake etc for tinkerers as well...

Cool. Let's take a well-known word prompt and rebrand it to pattern for no reason whatsoever, other than confusion. I know programmers love fancy labels but this is just forced
Honestly this looks cool but setting it up is a bit much. Like... I think the idea of this is cool, but by the time I've done all this setup I've more than likely forgotten what it is I was trying to do to begin with, and even when most of it's already installed I need to make sure several things are running etc. before this becomes essentially a link in the chain. That feels to me like it's going against the ease of integration the project is meant to facilitate a tiny bit. Like somebody else already stated in this thread, having a bit more of a ready-to-go solution, stick it in docker if you must, would probably help here.
We’re moving to go.
d0mine [0] already mentioned this [1] (and was apparently downvoted to oblivion for saying it), but there's already another Python project named Fabric that's been around quite a while now. [2,3] Downvote me too if you like, but it's not gonna change the fact that there's already a project using that name.

- [0]: https://news.ycombinator.com/user?id=d0mine

- [1]: https://news.ycombinator.com/item?id=40893147

- [2]: https://pypi.org/project/fabric/

- [3]: https://www.fabfile.org/

Can we not use the name of my favorite minecraft modloader for some AI project? Thanks.
Small AI boats sail a rough sea in search of the new world. They can sink, never find land, or get sherlocked by massive frigates if they do.

Everyone knows it would be better to fly, i.e. go up one layer and build generic public infra instead. But then they just float off into the obligatory-xkcd-verse: https://xkcd.com/927/

This incredibly insightful piece about AI infra and why it is a bad idea to build in the space was on here lately: https://news.ycombinator.com/item?id=40862436
Alan Kay on "AI", human augmentation and Doug Engelbart, June 2024, https://news.ycombinator.com/item?id=40868478

  One big distinction in the 1962 time period is that they thought of “machine intelligence” as being a kind of complimentary set of thinking tools that could be “symbiotic” to how humans were able to think. They were not at all thinking about something like a slave or a major domo, but something more like a research assistant or a “Memex” (the latter was a big influence on Doug’s thinking).

  In the very late 60s the “official AI researchers” started to think that something like “intelligent Greek slaves” were needed for the “Romans” (Americans), and became rivals to Doug’s notion of elevating human thinking rather than just elevating power. This was a bad idea then … and it’s a bad idea today.
I find it interesting that Engelbart, Kay, Jobs and even Elon have all been learned people who saw in AI a way of augmenting human capabilities, finding truth, giving kids new ways of learning about history and science, and tackling planet scale challenges.

Someone posted an old video here a few days ago of Steve Jobs giving a speech in the 1980s saying that his dream was to make the computer into a medium that allows you to have a dialogue with Aristotle.

Yet I’ve never heard anything like that from the guy who is center stage for AI today, Sam Altman.

Maybe Sam is a a visionary in his own right. Maybe be recognized that people just care a lot more about SEO, about not having to talk to customers anymore, and about the kids just being quiet.

There is Fabric Python library to execute shell commands remotely over SSH https://pypi.org/project/fabric/
When influencers code.
Pwned.
I vouched for this post. Not sure why it or previous submissions were dead. Seems interesting.