gojomo

The quality varies wildly across models & versions.

With humans, the statement "my tutor was great" and "my tutor was awful" reflect very little on "tutoring" in general, and are barely even responses to each other withou more specificity about the quality of tutor involved.

Same with AI models.

Further, "enough" training from another model's outputs – de facto 'distillation' – is likely to have similar effects as starting from a common base model, just "from thge other direction".

(Finally: some of the more nationalistic-paranoid observers seem to think Chinese labs have relied on exfiltrated weights from US entities. I don't personally think that'd be a likely or necessary contributor to Z.ai & others' successes, the mere appearance of this occasional "I am Claude" answer is sure to fuel further armchair belief in those theories.)

But of course local GPU processing power, & optimizations for LLM-like tools, all adancing rapidly. And these local agents could potentially even outsource tough decisions to heavierweight remote services. Essentially, they'd maintain/reauthor your "custom extension", themselves using other models, as necessary.

And forward-thinking sites might try to make that process easier, with special APIs/docs/recipe-interchanges for all users' agents to share their progress on popular needs.

But ultimately, your browser should have a local, open-source, user-loyal LLM that's able to accept human-language descriptions of how you'd like your view of some or all sites to change, and just like old Greasemonkey scripts or special-purpose extensions, it'd just do it, in the DOM.

Then instead of needing to raise this issue via an "Ask HN", you'd just tell your browser: "when I visit HN, hide all the AI/LLM posts".

But: if all developers did 136 AI-assisted issues, why only analyze excluding the 1st 8, rather than, say, the first 68 (half)?

No option has ever existed on iOS, despite the recent announcement assuring users "Local backups still exist". They've never existed for iOS users!

If you are alleging that Apple's own local Finder/Itunes backup of an iPhone includes Signal messages, that's not true, against reasonable user expectations, by Signal's own design choices.

Anyone who's counting on such local backups to save their histories is in for the same rude surprise I and many others have hit unaware:

https://www.reddit.com/r/signal/comments/1hgukpg/backup_and_...

Techniques can be arbitrarily old & common in industry, but still be a novel academic paper, first to document & evaluate key aspects in that separate (& often lagging) canon.

I used 'delving' in an HN comment more than a decade before LLMs became a thing!

https://news.ycombinator.com/item?id=1278663

And, counter to much intuition & forum folklore, it works for AI models, too – with analogous caveats.

But also: with regard to claims about what models "can't experience", such claims are pretty contingent on transient conditions, and expiring fast.

To your examples: despite their variety, most if not all could soon have useful answers answers collected by largely-automated processes.

People will comment publicly about the "vibe" & "people-watching" – or it'll be estimable from their shared photos. (Or even: personally-archived life-stream data.) People will describe the banana bread taste to each other, in ways that may also be shared with AI models.

Official info on policies, processing time, and staffing may already be public records with required availability; recent revisions & practical variances will often be a matter of public discussion.

To the extent all your examples are questions expressed in natural-language text, they will quite often be asked, and answered, in places where third parties – humans and AI models – can learn the answers.

Wearable devices, too, will keep shrinking the gap between things any human is able to see/hear (and maybe even feel/taste/smell) and that which will be logged digitally for wider consultation.

The new encoding can contain a FLOAT32 side channel on every character, to represent its proportional "AI-ness" – kinda like the 'alpha' transparency channel on pixels.

These inproved models do some valuable things better & cheaper than the models, or ensembles of models, that generated their training data. So you could not "just ask" the upstream models. The benefits emerge from further bulk training on well-selected synthetic data from the upstream models.

Yes, it's counterintuitive! That's why it's worth paying attention to, & describing accurately, rather than remaining stuck repeating obsolete folk misunderstandings.

But I'm not suggesting they'll advance much, in the near term, without any human-authored training data.

I'm just pointing out the cold hard fact that lots of recent breakthroughs came via training on synthetic data - text prompted by, generated by, & selected by other AI models.

That practice has now generated a bunch of notable wins in model capabilities – contra the upthread post's sweeping & confident wrongness alleging "Ai generated content is inherently a regression to the mean and harms both training and human utility".

I find it (very) vaguely like how a person can improve at a sport or an instrument without an expert guiding them through every step up, just by drilling certain behaviors in an adequately-proper way. Training on synthetic data somehow seems to extract a similar iterative improvement in certain directions, without requiring any more natural data. It's somehow succeeding in using more compute to refine yet more value from the original non-synthetic-training-data's entropy.

However, since then, a bunch of capability breakthroughs from (well-curated) AI generations has definitively disproven it.

guaranteed human output - anyone who emits text in these ranges that was AI generated, rather than artisanally human-composed, goes straight to jail.

for human eyes only - anyone who lets any AI train on, or even consider, any text in these ranges goes straight to jail. Fnord, "that doesn't look like anything to me".

admittedly AI generated - all AI output must use these ranges as disclosure, or – you guessed it - those pretending otherwise go straight to jail.

Of course, all the ranges generate visually-indistinguishable homoglyphs, so it's a strictly-software-mediated quasi-covert channel for fair disclosure.

When you cut & paste text from various sources, the provenance comes with it via the subtle character encoding differences.

I am only (1 - epsilon) joking.

Has a righteous, bossy tone that doesn't seem earned by case particulars or its (anonymous) author.

"Mozilla: Improve your messaging. Now."

And, while copyright prohibits some sorts of reproduction of copyrighted materials, it doesn't give rightsholders veto power over all downstream uses of legal copies.

Other impactful variants might be:

* senses whether another 'sibling' AirTag is present, if so, stays off. If not, waits X hours & then turns on.

* has its own motion sensor; only after X minutes of being stationary, it waits Y hours to turn on briefly

* has its own clock & (original-user-known) randomization seed; turns on at pseudorandom intervals the original user can predict

* low-power/low-bandwidth receivers so cheap & tiny now: could wait for national or even global unit-specific 'wake' request - perhaps even with parameters for duration/intervals – before powering-on AirTag portion

If instead users must use your web-served client code each time, you could subtly alter that over time or per-user, in ways unlikely to be detected by casual users – who'd then again be required to trust you (Tinfoil), rather than the goal on only having to trust the design & chip-manufacturer.