Exactly. The software engineering domain is huge. I could code just about anything when I was 16 years old after 2 years of intensive learning... It took me an additional 10 years to learn to design software that is secure, maintainable, efficient and scalable.

The problem with these kinds of discussions is that they act like experienced software engineers themselves don't bring domain expertise to software products.

So AI can easily replace the domain knowledge of software engineers but not of evey other profession?

Coding is not engineering but I'm glad that we will finally be able to prove that definitively thanks to AI. It's going to be a bumpy ride.

Any software engineer who has built software to solve domain problems in multiple industries knows that the engineering domain knowledge and systems thinking approach is far more difficult to attain than industry-specific domain knowledge... This is why there are software consulting firms which can work across multiple domains. Understanding the problem domain is not that difficult.

I'm in a similar boat. I don't like modern music. I was never a big music fan TBH though I did like a few really good pieces from my day. That said I never cared much for lyrics because I didn't find them relatable. I'm only interested in the tune... I like lyrics but only for the audio properties of the words; literally, I like the sound of human vocal chords.

The way I use Suno is sometimes I play Ukulele and discover a tune I like; I record it and generate a song from it.

I didn't take any music lessons. I'm 100% self-taught so my recordings are a little rough but the melody comes through and Suno polishes it up nicely and adds lyrics based on a topic I've been thinking about.

I find both the creation and listening aspects relaxing and therapeutic. I'm not a musician so Suno is the only way I could actually produce and finish a song. It's very clearly my melodies, my songs but it's enjoyable to hear them as a finished product. There is definitely an element of surprise, the lyrics are sometimes quite insightful and clever too and I can actually start appreciating the poetic aspect of music in a way which eluded me before.

I suspect that by the time most musicians finish refining and producing their own songs, without AI, they're probably tired of hearing it. Suno avoids that. It's a truly novel thing to be both a producer and consumer of your own music. Perfect for an introvert like me who can't relate to anyone except himself.

It's nice to see that some other people also like my pieces though I'm not trying to make a career out of it.

I got Netflix recently because I wanted my 4 year old son to have a bit of screen time. He is very active, loves to play outside, draw, paint and does a lot of different activities so I thought there's no harm in a bit of screen time... But YouTube was awful because he would end up watching creepy AI slop videos of extremely colorful 3D cars or airplanes or whatever... Very repetitive. Or sometimes there would be videos of colorful painted toys being washed with a monotonous voice repeating the same thing over and over and saying stuff like "Wow, it's big" in the most monotonous voice possible (and no, the toy was not even big)... There's something very creepy about hearing emotionally charged sentences being expressed with such dull apathetic tone and saying things that aren't even factually correct. Complete trash. I could feel myself getting brainwashed in realtime. No more YouTube. Definitely this AI slop should not be promoted.

The stock market should be illegal too. The stock market is just a prediction market at this point. What value does it create?

> The stock market is just a prediction market at this point.

Not really. But I can get your point.

> if an employee can predict specific search results to such a granular extent

It's not a prediction, it's a query in a tool/database that's supposed to hold counts.

There's no rigging here - it's just a summary of historical data being accessed and used before being public, in violation of confidentiality agreements. If he had just sourced the data without acting on it, the system would have still logged somewhere, but likely nobody would have looked at it. My understanding is that the last couple weeks before being released won't dramatically change counts unless an extraordinary event happens that drowns the last 11 months of data.

I reject the the idea that these types of people are needed. It's probably that most of the people in the CIA happen to be like that because they're power-hungry and they're just selecting their kin and justifying their choices as "right kind" because they narcissistically believe themselves to be the right type... They're probably the wrong type. Especially if they all share narcissistic or psychopathic traits; it's too many, it cannot work.

The company hires people who match the company’s desired culture.

If the people in the CIA who do hiring want the talent who are excellent at lying and compartmentalizing their ethics, then that’s what the organization becomes over a generation.

This is a great point. I think for coding, the wording of the MIT open source license makes it clear that copying and distributing the software is authorised on a small scale and it's very clear that the act of copying must involve a person.

It provides distribution and modification rights to "any person obtaining a copy of the software" and explicitly requires attribution for any significant parts.

Mass-ingesting the code with a script without any human even reading the licence is a very different kind of copying mechanism and there is no person involved... The contract was bypassed completely. A contract requires consent from both parties to be binding. When ingesting code into the AI training set, nobody even read the license. There was no agreement; neither explicit nor implicit... Because the consumer, a script, never read the contact for that specific project.

There was nobody present when the copying occurred; on neither side! It cannot possibly constitute an agreement between two parties.

> I think for coding, the wording of the MIT open source license makes it clear that copying and distributing the software is authorised on a small scale and it's very clear that the act of copying must involve a person.

I agree with “must involve a person. https://opensource.org/license/mit starts with (emphasis added) “Permission is hereby granted, free of charge, to any PERSON obtaining a copy of this software and associated documentation files (the “Software”)”.

That means it doesn’t give an LLM any rights. The way I see it, LLMs run (directly or indirectly) by a person can do stuff on their behalf, though, just as your CI pipeline can download and compile MIT-licensed software.

I definitely disagree with the “on a small scale” as the license continues (again, emphasis added) “to deal in the Software WITHOUT RESTRICTION, including WITHOUT LIMITATION the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software”.

The CI pipeline is different because for a module to end up as a dependency in the CI pipeline, it had to be explicitly selected by a person first to be included in the package file or manifest. There was intentionality and awareness that the software was included.

A person already pre-consented to the licenses of all the software which the pipeline downloaded. Big companies go through those dependency lists carefully already and remove those which do not meet their policies. This is a very intentional process.

> for a module to end up as a dependency in the CI pipeline, it had to be explicitly selected by a person first

I disagree. I think it’s entirely within the license to have your pipeline automatically pull in the latest version of a library, even if the new one happens to pull in a new MIT-licensed library (whether that’s a good idea and whether CI pipelines should, somehow, verify that code pulled in has an acceptable license are different discussions)

I also think it’s complete within the MIT license to tell a LLM that it can search for MIT-licensed libraries and use them without asking you.

This would be an extremely novel mechanism of copyright litigation and I doubt it would fly in an American court with its' emphasis on highly individualized legal rights and obligations. And, if it did get accepted by the courts, that's halfway to an even crazier argument: that the MIT license only allows individual distribution to known parties; i.e. no hosting the code on a website or seeding it on BitTorrent, because that's not "small scale" and doesn't "involve a person".

And the thing is, you point out the easy out on this for similarly licensed code... a giant list of authors and contributors that may have code included in the generated output. It's a win/win for everyone. The original authors get their acknlowdgement, and the AI company gets to bill the users of AI for all the tokens for that multi-gigabyte copyright disclosure file.

That's like saying you're not allowed to load the source code into an editor, because it's not a person. Or that you're not allowed to run a global search-replace on the entire code base, because it's a script and not a person.

Sorry that's nonsense. There's human awareness when ingesting MIT code into an LLM too. In both cases it's a human that says $ excute-global-replace or $ ingest-into-llm

Both operations require some degree of human awareness. What you appear to be saying is, a human can only use a limited algorithm to access this source code, not a sophisticated one. And where do you draw that line? Who should get to say what is too sophisticated?

Error: your algorithm is too sophisticated to proceed, please provide more human awareness, it's a critical difference.

If your LLM were to hack into Microsoft and steal the source code from an important project and inject it into your project without you being aware of it; wouldn't that make you liable if you then published it?

Unfortunately there is no way to agree to a license of a software you're using if you didn't read the license or if you're not even aware that you're using the licence. This is what's happening at the training stage.

If you say that awareness doesn't matter then it means you cannot stop AI from stealing any IP open source or not.

I think the main issue with LLMs is that there is no mechanism to stop them from stealing. Thus they are guaranteed to infringe on copyright to some extent.

Also, beyond copying and copyright, there is another problem that LLMs are also infecting the logic and expertise built into the project. This is a completely novel mechanism and needs to be treated as separate under the law. Else it would be the end of all IP.

> I think the main issue with LLMs is that there is no mechanism to stop them from stealing.

Well, sure there is—for the people running them.

If you're building training data for an LLM, you only use data that a) is firmly in the public domain, or b) you have a clear and documented legal right to use.

Page Manager Theme Editor Media Library SEO Settings Analytics Domain Manager Export Code Publish Button Realtime Preview Undo/Redo

I'm not too worried about it because the first segment of society is doomed to be 'good but never great.'

AI lacks the ability to identify greatness because it's trained on the output of the average person who also lacks this ability.

It's going to create a new elite class of people who have good taste and the masses who have bad taste. Many current elites will end up with the masses. They may retain their wealth on paper, but it will be a cheap, low-quality existence but they will be convinced it's luxury.

I think eventually, everyone will get what they want, but not everyone will get what they need.

Taste is subjective, authenticity is not. People in #2 want human created content, even if it's not as "good".

My definition of bad taste is; will be derivative. These people will consume variants of the same thing over and over, not realizing it to be the case. They will be narrow minded and predictable. They will be afraid of any other ideas which doesn't fit the acceptable pattern of their tribe.

I replied above, but they're already like this.

True but I don't think we've reached the limit.

This is why I advocate for people to spend some time outside in nature and try to do something different once in a while because it's so easy to get stuck in a really small bubble. Especially when you exist in an entirely man-made, soon-to-be fully AI-controlled environment. You may lose your ability to have novel thoughts.

Some people are already there but the range of thoughts seems to be narrowing.

My biggest fear is being caught up in such group for financial reasons and trying to navigate some kind of linguistic and conceptual minefield everyday. I already encountered a situation like that twice in my career. Very tense environment. Feels like you're in a brainwashing cult and have to pretend to be one of them; it's really hard to pretend to be ignorant of certain kinds of information when you don't know what the full range of forbidden ideas is. Saying the wrong things got me fired both times; differences in our mental conditioning created very subtle tension/discomfort between me and management.

They will tolerate people who are 'running a simpler program' than themselves but they will absolutely not tolerate someone with a broader programming. Hence you have to pretend to be narrow-minded which is hard to maintain. This is why I like remote work.

The elites already have bad taste. They aren't going anywhere without tax and financial reform and the return of political donation regulation.

We need to change the law to reduce regulations and introduce a principle of unlimited liability instead.

These things have to be fixed at the incentive layer. Self-regulation is the best form of regulation, by far.

Any discarded plastic found outside of a landfill should be tallied and result in a fine for the company which produced the plastic.

If a person is caught discarding a plastic bottle (littering), they should receive a fine as normal but the company should also be fined a portion of the liability.

The company knows very well that some percentage of the plastic they sell will end up in nature. They know this for a fact... And yet they choose to keep producing plastic packaging when cleaner alternatives exist; I.e. tin cans and glass bottles.

We've been completely brainwashed by the cult of 'limited liability'. It's a horrible idea. The words themselves tell you everything. Right there in front of your face. The liability is limited... It means the liability is externalized. This construct should never have been allowed.

Now with AI and the externalized harms which will result from it, this construct is more important than ever. Lives are at stake.

Once we start assigning partial liability to every harm. Eventually, we'll be able to collectively identify all of them; they will become so rare that they will stand out like a sore thumb.

Imagine; all the people who are currently messing up the political system with useless bureaucracy and toxic ideologies could actually be useful to society. Instead of identifying each other's genders, orientations and emotional triggers, they could be identifying social and environmental harms and holding companies accountable.