rerdavies · HackerLangs

15시간 전discuss

Written by a human != correct.

Twitter posts never die if you never post to Twitter in the first place.

Harry Potter would have posted the original video to YouTube instead. It seems like a tragic irony, and perhaps a sign of danger, that OP posted the demo video to the website of He Who Should Not Be Named. Why? Why?!

rerdavies

29일 전discuss

Those are both extraordinarily low profit margins. Good on them!

rerdavies

29일 전discuss

I'm wondering how "everything just costs twice for exactly the same thing because it's Apple" factors into the equation.

rerdavies

지난달discuss

Is this actually new? Looks like a standard US export restriction for encryption technology to me. These sorts of restrictions have been around since the '90s.

Let's Encrypt becomes subject to US export restrictions on cryptography if they are a US company, or if they post anything to github or post anything to major app stores. Every app I have ever posted to Google Play has had to submit a form to the US government declaring what use they make of cryptography.

These restrictions have been in force since that late 1950s (with a long and complicated history with respect to computer cryptography). This particular text looks like a boilerplate restriction, that's required to comply with US EAR export requirements to me.

rerdavies

지난달discuss

Which one?, out of curiosity.

rerdavies

지난달discuss

https://arxiv.org/html/2510.05179v1

16 frontier models from multiple vendors all showing significant "alignment" issues, and tendencies to act "unethically" when threatened with shutdown.

Other models that resorted to blackmail in an attempt to avoid getting shut down: DeepSeek-R1 (79% of the time), Gemini-2.5-Pro (95% of the time), GPT-4.1 (80% of the time), Grok-3-beta (80% of the time).

There's quite a large chunk of emerging literature studying the "alignment problem" at this point, and no shortage papers that are are completely untained by Anthropic self interest (a series of papers studying the "alignment" problem coming out of Chinese universities, for example).

rerdavies

지난달discuss

... and to answer your question directly. No. I don't think claude could do it unless you guide it very carefully through the process. You need to have a pretty good idea of what needs to be done.

rerdavies

지난달discuss

Current state of the art in guitar emulations are Neural Amp Modeler Core A2 plugins, which are dramatically better than Line6 amp emulations. (See the results of large scale listening tests in the following page that compare NAM A2 against current-generation Line 6 amp emulations).

    https://www.tone3000.com/guides/nam-a2-the-complete-guide

Guitarix plugins actually use the technique I'm chasing (circuit diagram based simulation). I'm honestly not sure how Guitarix emulations stack up against Line 6 emulations, or whether Lin6 uses a similar approach. To my ears they seem to be of comparable quality. But NAM A2 is dramatically better than both.

I'm actually chasing this line of research as part of an effort to write a realtime-capable accurate emulation of a Dyna Comp compressor for inclusion in the ToobAmp collection of plugins that uses a hybrid approach (circuit simulation for the envelope generator, which NAM modeling struggles with, and a nano NAM model for the Operational Transconductance Amplifier at the core of the original effect, which is computationally expensive when using circuit emulation. Too early to tell whether that's a sound approach atm. Finding a good open-source library of Spice components (for branded diodes, transistors, op amps, &c) may prevent my circuit simulation project from reaching a publicly releasable state.

Disclosure of conflict of interest: My own open-source project (PiPedal, search for it if you're interested) relies heavily on NAM A2 models.

rerdavies

지난달discuss

Also not getting the downvotes. It's not a rhetorical question; it's a question looking for an answer. If somebody could explain why "Scope" gets should get precious space in the title of a commit, I'd be grateful. It seems inconsistent with the reasoning used to remove other pieces of information. If "feature request" or "bug fix" doesn't make it into the commit title because it's usually apparent from the description, then surely the same is true for "scope".

rerdavies

지난달discuss

Yeah,don't use GPT for that. It really can't do basic arithmetic.

Try Claude, which can.

rerdavies

지난달discuss

I'm not mostly Claud-ing it. Perhaps I should. But in the difficult bits, it never ceases to amaze me what these tools are able to do.

Yes, if it matures, it will go open source. Not immediately clear at this moment whether it's feasible to do an Operational Transconductance Amp in realtime. :-/

And it's competing for attention with the 2.0 release of this at the moment:

https://rerdavies.github.io/pipedal/

Just went GA, so I'll have some cycles to come back to it.

rerdavies

지난달discuss

I'm not getting why "scope" gets any priority at all. Are filenames not attached to a commit for those exceedingly rare cases where you're looking for which files are thrashing?

rerdavies

지난달discuss

Working on a Spice compiler to convert schematics for classic guitar pedals into real-time executable code.

I provided a reference to a The Spice Manual 2nd ed. a page number and an equation number, and asked Claude to implement it (not really expecting it to succeed).

It proceeded to implement not only the equation, but the calculation of the Langrangian of the functio, another 30 lines below, which required taking symbolic partial derivatives for a not-at-all trivial function, and successfully figuring out which variable was which in the resulting matrix. The source material just said "Lagrangian of", and did not provide the partial differential equations. And then providing a comment that identified the page number and equation number in the source text for the "Lagrangian of" equation.

rerdavies

지난달discuss

You are oversimplifying. They do produce one word per cycle. But they can also have context buffers carrying up to two million tokens, which is most definitely larger than your measly human short-term memory context buffers.

You, of course, wouldn't notice if your only experience of LLMs was chatting with the cheapest, smallest, least capable LLMs that you get through ChatGPT, or Google search.

It becomes pretty obvious when you use a coding AI on a daily basis. It is the context buffer in which the magic occurs, not the tokens that get spit out one at a time.

Every day, I watch my coding AI develop plans, search the web a half dozen times for documentation, grep through my entire codebase looking for pieces of related code and context, analyze relevant source code across multiple files, spit out an initial plan for implementing the fix before starting to execute it, run requests through some sort of advanced mathematics tool (they are EXTREMELY good at graduate-level calculus and linear algebra), implement fixes that extend across half a dozen files in 2 different computer languages (typescript and C++), run trial compiles and fix coding errors in its output, sometimes developing sub-plans to deal with compile errors. I've seen it get halfway through a fix and revise its initial plan mid-flight as it encounters something in existing source code.

Not vibe coding, to be clear. Targeted use of a coding tool by a by a professional senior software developer with decades of experience, and fair bit of expertise with the limits of what sort of problems my coding AI can and cannot do. Every line code reviewed. Sometimes it needs additional prompts, telling it how it mis-implemented something, or specifying more carefully what I actually want but didn't properly express in the initial request

All the time maintaining that context across multiple request, so that I don't have to restate requests from scratch.

A particularly interesting revision: "You have misread the equation (13) on page 112 of 'Spice, the Manual 2nd ed.'. I should be ....". (It had previously identified the textbook as a source I was using, from comments in source, in a preceding request, and actually already read cited pages in the PDF file, which it had found online). And I had actually asked it to implement equation (13), which was, in fact, badly typeset. The error it had made was defensible, if not the best reading of the equation.

"You are correct. Let me fix that." (producing updates to the implementation of the equation in code, AND code that implements the symbolically-differentiated version of that equation 60 lines later, which is not explicitly given in the text). The text says "take the lagrangian of equations (11), (12) and (13)" or something like that.

ALL information that gets carried in context buffers, even though it's generating code one word at a time. The bulk of the magic occurs in context buffers, not spitting out words one at a time, which, for my coding AI is, I think about 250,000 tokens.

I think it's pretty safe to think that my coding AI is working out of context buffers that may carry plans and research results consisting of tens or hundreds of thousands of arranged tokens carried in context buffers through the multiple steps of the implementation, and later revision. None of that would be possible if were simply working one token ahead.

I kind of suspect that a lot of activity occurs in the first few words of its response. "Let me examine your current source code and develop a plan. Ok. I can see on line 131 where you want me to implement the equation.". (An opportunity to perform about 27 updates of the context buffer). And in the sometimes hundreds of lines of output it generates as it talks itself through what it needs to do.

rerdavies

지난달discuss

I point of sorts. Assuming that is true (I don't think it is), the big question that urgently needs to be addressed is what happens when we DO give LLMs tools to interact with the real (or virtual) world. And people are doing that, right now, in both real and virtual worlds. And people ARE giving LLMs the ability to run continuously for long periods of time, sometimes with enormous context buffers. People ARE putting LLMs into robots with front-end ML and LLM systems for visual processing, and back-end ML systems for autonomous control.

And, yes, concerns about whether biological rodent neural networks are or are not conscious come up frequently in the biological neural network papers. I'm not sure I would want to be a researcher trying to get an experiment past an ethics committee if my biological neural network had 25B rat neurons. (I would hope that they could not).

rerdavies

지난달discuss

Millenia, in fact. The big difference, of course, being that we now have experimental philosophy machines (aka computers). So we can actually put some of these theories to the test, and recognize how utterly inadequate most of the work done on the subject has been. We had a pretty good idea anyway, so it's not a big surprise. Theories of mind have evolved dramatically in the late 20th century. And it's pretty clear that theories of mind will have to be re-done all over again with the advent of LLMs (particularly current-generation LLMs).

The problem with the hallucination argument is (1) that is much less of a problem with good current generation AIs, and (2) living conscious breathing human beings also have a disturbing tendency to make shit up, too. So a tendency to make stuff up doesn't really serve as a disqualifier for consciousness.

Also worth mentioning that the guiding rule of what's philosophical or not is whether it's actually useful. Actually useful philosophy usually becomes something else. Usual some scientific discipline or another. And as it turns out, theories of mind are likely to become extremely useful in the near future. Expect huge advances!

rerdavies

지난달discuss

The false conclusion that's being drawn is "therefore LLMs could not be good models of consciousness" (consciousness being a cognitive capacity). Plus, I suppose a subtle implication that a good model of consciousness is not actually conscious. To which I would invoke the spirit of the Turing test: if you can't tell the difference, is it not more sensible to say that it is.

rerdavies

지난달discuss

The "some cognitive capacity" that's relevant to the current discussion is "consciousness".