anileated

> There are many cases in which I already understand the code before it is written.

Typing speed is your bottleneck?

Replacing direct input interfaces with LLM chatbots is not unlike “upgrading” from a modern videogame (be that Call of Duty, Disco Elysium or Dwarf Fortress) to a retro text-based adventure. And if you have a direct input interface, why do we need an extra expensive layer of non-determinism?

> I would say, this is not about the final product, but a way of creating a product.

It is the same logic as not wanting to use some blockchain/crypto-related platform to get paid. If you believe it is mostly used for crime, you don't want to use it to get paid to avoid legitimizing a bad thing. Even if there's no doubt you will get paid, the end result is the same, but you know you would be creating a side effect.

If some way of creating a product supports something bad (and simply using any LLM always entails helping train it and benefit the company running it), I can choose another way.

If your bottleneck is not typing speed, you'll be fine.

There seems to be some confusion. I wouldn't call conservative politics as a whole fascist, that's your choice of words. I doubt that "anti-AI progressive" is a thing too.

> The rest is arguably not objectively bad but more personal or political opinion from your side.

Nothing is objectively bad. Plenty of people argue that gambling should be legal if anything on the basis of personal freedom. All of this is a matter of personal choice.

(Incidentally, while you are putting people in buckets like that, note that one person very much can be similtaneously against gambling and drug legalization and be pro personal freedom open-source libertarian maximalist. Things are much more nuanced than “progressive” vs. “conservative”, whatever you put in those buckets is on you.)

If your argument is that there are more jobs that require morally dubious developments (stealing people's IP without licensing it, etc.) than jobs that don't, I don't think that's news.

There's always more shady jobs than ethically satisfying ones. There's increasingly more jobs in prediction markets and other sorts of gambling, adtech (Meta, Google). Moral compromise pays.

But if you really think about it and set limits on what is acceptable for you to work on (interesting new challenges, no morally dubious developments like stealing IP for ML training, etc.) then you simply don't have that FOMO of "I am sacrificing my career" when you screen those jobs out. Those jobs just don't exist for you.

Also, people who tag everybody like that as some sort of "anti-AI" tinfoilhatters are making a straw man argument. Most people with an informed opinion don't like the ways this tech is applied and rolled out in ways that is unsustainable and exploitative of ordinary people and open-source ecosystem, the confused hype around it, circular investment, etc., not the underlying tech on its own. Being vocally against these matters does not make one an unemployable pariah in the slightest, especially considering most jobs these days build on open source and being anti license-violating LLMs is being pro sustainable open-source.

> From purely business and career perspective, being anti-AI will be a self-own

From purely business and career perspective being anti-blockchain/NFT/online gambling/adtech/fascism (at least for now in US)/etc. is a self-own, too.

I'm sure everybody making a choice against that knows it.

Thankfully purely business and career perspectives don't dictate everything.

Hat tip on both your (new?) domain name and your username.

This is not under that “all discussion” because the topic is not the same.

If the title of the thread is ”Microsoft restructuring”, and I care about them becoming for profit, I am not clicking on that thread, the headline is not interesting.

I come here pretty often and have not seen a proper discussion about them becoming for profit in the past month. Last it happened was a whole year ago, and then it was just announced and was not confirmed.

Why is it relevant for the topic of becoming for profit that someone discussed specifically the Microsoft topic recently? Unless the topic of becoming for profit is intentionally kept under wraps.

That, of course, is not about its transition to a for profit, which is what this news is about.

Probably all of the ML foundation like transformers which OpenAI used to create its chatbot was originally developed at Google.

OpenAI is for-profit: https://www.theguardian.com/technology/2025/oct/28/openai-fo...

That’s why everybody uses context to understand the exact meaning.

The context was “when would an AI agent doing something it’s not permitted to do ever not be bad”. Since we are talking about a tool and not a being capable of ethical evaluation, reasoning, and therefore morally good or bad actions, the only useful meaning of “bad” or “wrong” here is as in “broken” or “malfunctioning”, not as in “unethical”. After all, you wouldn’t talk about a gun’s trigger failing as being “morally good”.

Bad: 1) of poor quality or a low standard, 2) not such as to be hoped for or desired, 3) failing to conform to standards of moral virtue or acceptable conduct.

(Oxford Dictionary of English.)

A broken tool is of poor quality and therefore can be called bad. If a broken tool accidentally causes an ethically good thing to happen by not functioning as designed, that does not make such a tool a good tool.

A mere tool like an LLM does not decide the ethics of good or bad and cannot be “taught” basic ethical behavior.

Examples of bad as in “morally dubious”:

— Using some tool for morally bad purposes (or profit from others using the tool for bad purposes).

— Knowingly creating/installing/deploying a broken or harmful tool for use in an important situation for personal benefit, for example making your company use some tool because you are invested in that tool ignoring that the tool is problematic.

— Creating/installing/deploying a tool knowing it causes harm to others (or refusing to even consider the harm to others), for example using other people’ work to create a tool that makes those same people lose jobs.

Examples of bad as in “low quality”:

— A malfunctioning tool, for example a tool that is not supposed to access some data and yet accesses it anyway.

Examples of a combination of both versions of bad:

— A low quality tool that accesses data it isn’t supposed to access, which was built using other people’s work with the foreseeable end result of those people losing their jobs (so that their former employers pay the company that built that tool instead).

Hope that helps.

An LLM is a tool. If the tool is not supposed to do something yet does something anyway, then the tool is broken. Radically different from, say, a soldier not following an illegal order, because soldier being a human possesses free will and agency.

I got an impression that you can, but I have not dug into developer reference deeply enough to cite it for sure.

True, it was incorrect for me to say they already do parallel execution. However, when parallel execution is a special case of speculative execution, the security concern I meant to highlight still applies, doesn’t it?

Yup.

Since the author of the app mentioned reverse engineering, analyzing audio is a way that immediately came to mind. It should be quite precise, too, only at the expense of extra CPU cycles.

I did not imply that there is no API to get head tracking data (even though Google search overview straight up says that). It’s mostly a thought experiment. Kudos for digging up CMHeadphoneMotionManager.

> Apple does not secretly analyze sine waves to infer head motion.

Duh. The mechanism I described hinges on Apple being able to track head movements in the first place in order to convert that virtual 3D scene to stereo sound.

> I've always wondered if any CPUs have tried to reduce the branch penalty by speculatively executing both ways at once in parallel

They already do it (edit: they don’t). It is difficult to get security right, however (see https://en.wikipedia.org/wiki/Spectre_(security_vulnerabilit...).

Put a sine wave emitter (or multiple) on the scene. Enable head tracking. Analyze stereo sound at the output. Mute output. There you go: you now can track user’s head without direct access to gyroscope data.

I wrote:

> There is nothing in the word “theft” that implies depriving someone of physical property.

It is always depriving of something. Just not always of physical something. The rest of my comment goes on about depriving of non-physical things.

> It's pretty metaphorical to describe such things as property that can be stolen.

Stealing is a pretty wide term meaning deprivation of a good, asset, property, service, etc. If it means deprivation only of physical goods at some point in time, sure; this is 2025 outside now. Theft of physical goods is so first millenium.

> Theft (from Old English þeofð, cognate to thief) is the act of taking another person's property or services without that person's permission or consent with the intent to deprive the rightful owner of it

> for over a thousand years theft has mean taking something from someone permanently

Nothing in your comment is in contradiction to mine, or suggests that the word has been “overloaded”. That’s what theft is. Intellectual property is property, trade secrets are property.

There are intangible kinds of property and assets that are more valuable than “real property”. Trade secrets is an obvious one, being a special case of intellectual property. It would be absolutely irrational for legal system to classify theft of iPhone as criminal and theft of IP as civil, when the former cost $700 and monetizing the latter could finance creator’s entire life.

There is nothing in the word “theft” that implies depriving someone of physical property.

Theft of private data deprives the owner of privacy. Theft of corporate secrets deprives the company of competitive advantage (and if not prosecuted, economy at large of incentives to innovate). IP theft deprives IP holder of ownership claim (and if not prosecuted, arts at large of incentives to create). Identity theft deprives the identity holder of whatever access to their identity provided to them. This can be continued infinitely.

These scenarios are not the same, and using “theft” for all of them is not precise. However, it is 2025 and in developed countries this sort of crime happens more often than basic theft of physical property, and the detriment from it is often much, much more severe than from basic theft of physical property. (I am sure I don’t need to explain how depriving IP owner of ownership claim can cost the original creator much more than depriving them of some single physical asset, both literally financially and in terms of psychological damage.) It’s therefore important to have a short, mainstream, easy to understand and non-legalese term for these scenarios.

Without any suitable mainstream term the word “theft” is a good enough intuitive approximation—if anything, it’s a bit too mild of a term.

Again, I agree it is not the same as theft of physical property. In fact, I point out that it is worse. Yes, unfortunately English language lacks a better word.

That’s what LLM would say…

I think it is safe to assume that more generous (for them) interpretation is the one that will be used by any big company. My link:

> You grant us and our legal successors the right to store, archive, parse, and display Your Content, and make incidental copies, as necessary to provide the Service, including improving the Service over time.

> parse it into a search index or otherwise analyze it on our servers

This is an “AI” platform now. “Improving the service” means that. “With your consent” means you have accepted the ToS (which by the way can be changed at any point and your continued use of the service means you consent to it).

The link you posted is not ToS.

https://docs.github.com/en/site-policy/github-terms/github-t...

(The link I posted is also not the proper ToS, it is more of an abridged version. They made the actual ToS somewhat hard to find and I cannot be bothered.)