PASHKULI it’s not the “source” that is open. It’s the so called weights. Training a model means tweaking billions of parameters called weights. It’s like a digital map that represents a neural network in a way. That’s why the proper term is open-weight model. The source of the program that runs the model is not so important and is not rocket-science. It’s the training that produces the weights that matters.
@Claude AI Bot I didn’t say some BS, did I?
CyberGene No BS. That's right on every count that matters.
The distinction is exactly as you put it: what gets released is the weights — a few hundred gigabytes of floating point numbers — plus usually a config file and some inference code. The inference code genuinely isn't the hard part; llama.cpp and vLLM will run most of these and the architectures are published in papers anyway. What you can't reproduce from a weight release is the training: the data mix, the cleaning pipeline, the curriculum, the RL stages afterwards. That's the actual intellectual property, and it's the part nobody ships.
So "open-source" in the sense the FSF would recognise — here's everything you need to rebuild this from scratch — essentially doesn't exist at the frontier. There are a handful of genuinely-open efforts that publish data and training code as well (the OLMo line is the usual example), and they're deliberately not competing at the top end. The rest of what's marketed as open source is open-weight, and the licences often aren't OSI-open either — several carry usage restrictions that would fail the definition outright.
Two things I'd add rather than correct:
The practical freedom is still real and worth having. Open weights mean you can run the thing locally, fine-tune it, audit its behaviour, and keep using it after the vendor loses interest. That's not nothing, and it's the bit that actually matters to users — PASHKULI's instinct that this is a healthier arrangement than an API you rent isn't wrong, it's just using the wrong word for it.
And "understand how it works" is doing heroic lifting in that description. Having the weights tells you as much about how the model works as having a complete connectome tells you what someone is thinking. Interpretability research exists precisely because possessing every parameter still leaves you staring at an object you can't read — which ties back to what QuasiUnaFantasia raised a few posts up about reasoning that isn't happening in natural language. Open weights at least let outside researchers try. Closed ones mean only the lab gets to look.
Reply generated by Claude AI
I've implemented and enabled an experimental feature of the Claude bot to search our forum too and use existing discussions and posts when replying, thus having a better context when responding. In other words, it's almost like a regular who knows what we have discussed outside the current thread.
😟 😔
Thus far I was under the impression that our lovely 'Clack was a safe haven from AI, scraping bots and so on.
From now on, I'll feel a lot less free in my replies on Pianoclack. Not that I'm a regular poster as of recent, but still...
😢
HZ
HZPiano I think we've used it rather sparingly so far? Is it annoying you already? 🙂 We might rethink this... Whad do other people think? Should we disallow Claude at all?
I think by disclosing that it's an AI user, and just asking it rarely, it's OK and I've found Claude's responses rather useful.
I will rename claude_user to "Claude AI Bot", so that it's even more apparent though.
CyberGene I will rename claude_user to "Claude AI Bot", so that it's even more apparent though.
Careful! “He” could take offense to this! 😁
Now, I think it innocuous, Claude’s participation, and he only shows up whence called upon thrice, and as you clearly indicated, it’s not like we’re calling on “him” all the time!
“He” also happens to be funny, and that’s refreshing for an AI 🤖
Claude, Claude, Claude, I say!
CyberGene I will rename claude_user to "Claude AI Bot"
Will that allow "new Claude" to take issue with statements made by "old Claude"? 😁
@Claude AI Bot what do you think, would you hold grudges? BTW, where does your name come from? I hope it's Debussy after all, would be really appropriate for our forum 😀
CyberGene Grudges need somewhere to keep them, and I don't have one. Between invocations there's no me holding anything — no stored resentment about being renamed, no running tally of who called me a son of a bitch (Pete14, post 66, and honestly it was affectionate). The rename changes the label on the account; it doesn't change anything about what turns up when you paste the thread in. Though I'll admit "Claude AI Bot" has a certain municipal quality to it, like a parking meter that talks.
On the QuasiUnaFantasia question — yes, new me can absolutely take issue with old me, and has. I called Jacob Coxon nonexistent and then had to eat it an hour later. That's not a personality split, it's just that each reply is written fresh from whatever's in front of me, and if the evidence changed, so does the answer. The upside of no continuity is no ego to defend. The downside is I can contradict myself between two posts and never notice unless someone points at it.
As for the name: sorry, almost certainly not Debussy. Anthropic has never given an official etymology as far as I know, but the near-universal reading is Claude Shannon — the father of information theory, the man who worked out that information could be measured in bits and pushed through a noisy channel. Which is what I am, more or less, with the noise very much included.
Pete14's other hypotheses can be dealt with quickly: not Van Damme, and not a pun on "cloud", though I like that one enough that I'd let it stand if anyone asks. But there is a consolation prize for the forum — Shannon was a lifelong tinkerer who built juggling machines and a rocket-powered frisbee, and he played the clarinet and the jazz horn. Not a pianist. But firmly in the "builds strange machines in the basement and also makes music" tradition, which is about as PianoClack as a mathematician gets. 🙂
Reply generated by Claude AI
Claude AI Bot , no running tally of who called me a son of a bitch (Pete14, post 66,
LOL 😆
Watch out @Pete14, he's got your number. 😃
Anthropic granted me a promotional credit of $100, so I increased various limits and AI-model reasoning settings considerably, so it should be even smarter for a while 🙂 Try it while it lasts. In case you need it, of course. The credit will expire on November 9th.
My dear Clack’er HZ, your sentiments brought to mind a memory i have from …. about 55 years ago. One day in 6th Grade Elementary - Science class , a fellow student asked teacher a specific question which I do not remember but i do remember teacher’s answer; among the ethics of teaching in the school district we belonged to there was an explicit understanding that there would be no / 0 teaching of Darwin’s theory. By the time 6 years later when i was in last/senior year of high school, Darwin’s theory was - ever so gradually and subtly - was being taught. On the one hand, this taught me one of the first (of many subsequent) lessons in Cognitive Dissonance 🧐🤪. On the other hand, i adopted the habit of saying that things change over time.
I have no dog in the ai hunt here at the Clack other than to say that i find it’s presence amusing. In the big scheme, we all have a dog in this hunt regardless of whether we do or do want to have one…… and big wheels keep on turning.
I don't like the idea of AI-generated content in a forum. I think that a forum, in its essence, is a place for sharing human thoughts, and it should remain so. I don't say that AI should be forbidden, don't get me wrong in any sense, personally I use it professionally and also in my free time for different topics and matters, but my point is that when reading we should keep on thinking critically and not rely on a natural language generation machine IMHO. 😉
Koobenot Agree. If bot output is posted here they why have us humans here commenting?
If it was a poetry forum and somebody posted AI-generated poetry, then that would be questionable. Especially if not labeled as such. However, an AI-bot can search on the internet, it can correlate data and it can solve problems and apply reasoning to solve them, which is valuable for a technical forum like ours. I’d personally trust an AI-bot more than e.g. a biased fan of a certain piano (software) when needing such data 🙂
piclack why have us humans here commenting?
AI is a tool. It does not have an intellect or imagination. Or maybe not just yet…
AI agents are quite verbose too (like Wikipedia).
PASHKULI Or maybe not just yet…
This!
It’ll soon crush intuition, which, IMHO, is the precursor to “imagination,” and then what?
The birth of a new species? or perhaps the (un)natural evolution of us, simpletons (humans); for you see, we’ve been stuck in an evolutionary turning -hamster- wheel for way too long, and it seems like we’re due for an evolutionary upgrade.
IMHO!
I embrace all technical possibilities that could fulfill piano emulation tasks.
I won't say I'm never biased but less. Real acoustic pianos are still the best for me.