HNHacker News
TopNewBestAskShowJobs

dTal

18,127 karma · joined June 2, 2013

submissionscomments
dTal··on Kentucky man becomes the first humanist granted religious exemption from AI
Well hang on, "just work" implies an objective standard of correct vs incorrect. No such standard exists. I'm sure that there will be legal AIs that reliably do... something... but the whole point of the legal system is to allow space for humans to argue their case. If an AI "does legal" deterministically, without human intervention, how can it be said to be "working" if its very use defeats the purpose of the system it has learned to navigate so expertly?
dTal··on Police Are Using AI to Pull over Drivers for 'Suspicious' Behavior
Your example evidence set is actually quite robust by US justice system standards. There are countless cases of innocent people getting sentenced to death for e.g. being the one to report the murder, often in the face of strong circumstantial counter evidence.
dTal··on BBC edits out pride pins on University Challenge in 'Orwellian' move
>a BBC spokesperson said: “In line with our editorial guidelines, we advise that no badges, flags, and emblems or symbols are worn or visible during filming, along with clothing with obvious branding or slogans.

Oh sure. In a few weeks the BBC will be drenched in poppy pins. I'll wager my left nut they won't edit those out.

dTal··on Securing our frontier model training environment
>the sandboxes themselves may need to be kept up to date

The sandboxes don't need internet access to be kept up to date. They just need to be writable by something with internet access. There's a million ways of doing this. Run things under a hypervisor. Have a PXE boot server in your data center that serves an "updater" boot image, after performing host attestation. Or if you really want them to "update themselves", physically plug em in for the update, physically unplug them the rest of the time, and don't run updates during agentic testing.

>or may need to be monitored

The sandboxes don't need internet access to be monitored, they just need to be readable by something with internet access. Like a hypervisor, KVM. or a webcam pointed at the screen, in roughly ascending order of paranoia.

>or the sandboxes may need to updated

Did this one already! See above

>or they may be connected to specific services

That's just "not airgapped" rephrased.

This is just off the top of my head and I'm no infrastructure engineer. OpenAI saying "it's too hard" is a major cop out. The real reason is that they show no signs of actually giving a shit about AI "safety".

dTal··on We are all living inside of an obscure sci-fi film from 1970
>The iRobot movie is like a parochial villager's imagining of what Asimov's books might contain before they have read them.

Eh. It's a surprisingly direct adaptation of "With Folded Hands" by Jack Williamson. Even the descriptions of the robots match.

dTal··on New plastic pollution discovered: 'Plasticorals' coral rubble fused with plastic
Oh well as long as there's a tardigrade or two left that's all fine then...
dTal··on New plastic pollution discovered: 'Plasticorals' coral rubble fused with plastic
There isn't infinite time left in "eventually" - maybe a billion years, out of 4.5 billion passed so far. A billion years is about how long it took to go from "multicellular colonies" to "concept of an animal". So no, not guaranteed to bounce back from mouldy radioactive slurry.
dTal··on Every AI model has the same "controversial" take on humans
If you had been asked, in 2016, to describe the behavior of a "probabilistic database", I very much doubt you'd have described the behavior of 2026 LLMs. They are deeply surprising in many ways.
dTal··on WorldTUI: I trained a 360M model to pretend to be a terminal
Hmm, so mosh augmented with this?
dTal··on In Hilbert space, all things are quantumly possible
Indeed such a strange configuration of matter would arguably no longer be "you", but I think an even more immediate point of philosophical incoherence is that it would certainly not be a cucumber. [Cucumber != cucumber] => false, QED.

If talking cucumber was too pedestrian for you, we could add stipulations until your intuition collapses. Perhaps, starting next week, the Earth begins to spontaneously turn into raspberry custard on Tuesdays. Perhaps the entire universe turns to custard! Just think of it - billions of lightyears of custard in every direction. Whoops, did I say billions? Too grounded! Let's make it a googolplex. And a single lone piece of rhubarb, floating somewhere in the middle of it all...

...sorry, couldn't resist. Joking aside, while we can quibble about the boundaries of the possible, the fact remains that there are boundaries - at least, if we're still talking about known physics. No evolution of the universal wave function will lead to a world where pi suddenly equals 3.

dTal··on What shall I teach my son?
They function just fine as either 1) subsistence farmers, or 2) beneficiaries of an enormous machine designed largely by people who knew calculus. You used to be able to function fine as 3) subsistence hunter, but then members of group 1 destroyed most of the wild ecosystems, while group 2 paved what was left and turned it into luxury apartments.

For better or for worse, humanity has chosen the path of gods. We'd all better start thinking like gods.

dTal··on In Hilbert space, all things are quantumly possible
Many Worlds Hypothesis is the extrapolation of taking Hilbert Space at face value, and dumping concepts like "wave function collapse" which privilege a particular future as the "real one that happened".

There's a possible future where you're President, but no possible future where you're a talking cucumber. This intuitive fact gets overlooked in pop culture explorations of Many Worlds, such as Everything Everwhere All At Once, which has ridiculous parallel universes which obey the rather more relaxed rules of "anything the writers can think of that looks silly", rather than "anything which can conceivably happen".

dTal··on DeepSeek banned from U.S. federal civilian agencies under new DHS order
That would have to be a stuxnet-level targeted sting operation, with a general purpose LLM exhibiting some ridiculously specific behavior in ridiculously specific circumstances, in such a way that nobody catches it in testing (and none of the other zillions of users of that general purpose LLM notices anything odd). If you're so paranoid that you do a blanket ban rather than just saying "hey guys maybe double check before using Chinese software in weapons we plan on maybe pointing at the Chinese", frankly LLMs should be the least of your worries, supply chain wise. Every US government employee uses a phone made in China (because they all are), and the president doesn't even update his, so color me skeptical that this has anything to do with security posture.
dTal··on DeepSeek banned from U.S. federal civilian agencies under new DHS order
The orders bans the use of open weights provided by DeepSeek too. It bans the use of the open source inference engine DeepSeek released today.

Sure. Wayyy more secure to shovel federal data into OpenAI /s

dTal··on Iosevka is an open-source, sans-serif and slab-serif monospace font
My personal favorite. Very underrated font, judging by this thread.
dTal··on Worlds more habitable than Earth?
Maybe better to start with a woman...
dTal··on A look at the math that OpenAI claims computes faster Fourier transforms
They mean physical distance, not "address distance".

And in the general case, memory reads do depend on distance. Hence:

> In practice there are multiple levels of cache, and cached reads will be much, much faster than main memory reads.

This is why we have the cache levels in the first place. And the hierarchy continues: L1, L2, RAM, block of spinning rust in the closet, "the cloud" somewhere... There is only so much storage you can fit close to core! More = physically bigger = physically further away.

Sure, within a level might feel like constant time. But that's because the optimal case for that level is slowed down to match the suboptimal cases. The step discontinuities in access time are caused by our inability to engineer a system to manage this tradeoff continuously, not a fundamental law.

dTal··on Lonely Reality of Self Hosting
The problem here is that from his family's perspective, it isn't self hosted - just hosted by a less powerful, likely less competent feudal lord than they're used to.

>I had this vision of my entire family using the applications I installed.

>occasionally asking me to install something new

If the model doesn't give them sovereignty, why should they care?

dTal··on Using AI for just 10 minutes erodes your ability to persist at hard things
Calling scipy.spatial.Delaunay feels about the same level of abstraction as telling an LLM "write me a routine to calculate Delauny triangulation", except the LLM can do it in any language, and I can trivially tweak the algorithm if I need to. If anything, using LLMs to generate code feels like going down a level, because I can now be bothered implementing things from scratch where previously I would simply have used a library.

You can't convince me that shepherding an LLM to re-implement a library function involves less "engineering" than just... calling the function.

dTal··on AI models can provably hide arbitrary capabilities
SaaS inference services can just spy on you passively, the old fashioned way; no need to bake nefarious behavior into the model itself, except perhaps to make it 'curious'.
dTal··on Whatever became of holography? (2011)
Well, yes, but actually, no...

"Hologram" is one of those words that has a wide variety of overlapping meanings. You can be a strict prescriptionist about it and insist that a real hologram involves laser light, interference patterns, 3D scene replication, phase preservation etc, but there will always be something that doesn't quite meet your definition that nevertheless is called a "hologram" by large numbers of people, even expert optical scientists:

A digital holographic stereogram prints an array of hogels (holographic pixels); the hogels themselves are holograms, but are mutually incoherent across the plate. When implemented to the highest level, the result is perceptually and behaviorally indistinguishable from an analog hologram.

A rainbow hologram (such as are found on credit cards) has single parallax only - it certainly does not capture all the properties of any particular electromagnetic field. No purity here, it's just a nifty tech for making things look 3D that happens to exploit wave interference. And everyone calls those holograms.

A scratch hologram is, to a first approximation, an extremely coarse version of a rainbow hologram, made by hand by engraving curved grooves in a reflective material. Now we're away from lasers and wave interference, and for this we sacrifice color - but geometrically we are still dealing with directional microreflections on a surface. Hologram? Depends who you ask.

Lenticular holograms use refraction rather than diffraction to achieve essentially the same imaging model as a rainbow hologram, but in a decidedly less finicky way, at the expense of some resolution. The same imaging data could be used to make a lenticular or a rainbow hologram. No lasers, no diffraction. Hologram? Depends who you ask...

My advice, as someone who has worked with holograms for a decade, is to give up. Saying "AKTUALLY Tupac/Abba isn't a hologram" gets old after the 1000th time. In the real world, "hologram" means "intangible optical simulacra", regardless of how it was achieved. If you want more precision than that, you have to specify the actual technology - Denisyuk, rainbow, lenticular, etc...

When someone asks "can you make a hologram of this", you don't lecture them on optical physics. You say "sure, you want the expensive kind or the plastic kind?"

dTal··on Whatever became of holography? (2011)
>Let people send you their 3D models or 3D scenes and you'll send them a hologram of it.

Oh hey, I have worked in exactly this business (taking people's random 3D models and bullying a rendering pipeline into feeding a digital holographic printer). It's hard, for many reasons. Mainly it's that holograms are expensive in materials, labor, and capital outlay (adding up to several k just for something the size of a piece of paper), and mass production of digital holograms isn't really possible without massive further outlay - writing them is time consuming, and copying them is fraught. It's like being a graphic design / print workshop, except everything is way more fiddly and expensive and nobody wants your product.

Analog holograms are a bit more tractable. But that's even more niche on the consumption end.

dTal··on An Anthropic AI model sent a false homicide tip to the police
I mean, it was spam...
dTal··on Why the AI future will be awesome
Biggest danger of LLMs as they currently fit into society: the radical consolidation of information flow through corporate-controlled bottlenecks that make current generation algorithmic social media (rancid as it is) look like child's play. When people no longer "search" for things, but simply "ask ChatGPT", the lever of social control is nearly absolute.

(Of course, all this is up in the air. If LLMs become commodities, the situation becomes very complex; the information flow doesn't get quite so consolidated, but it still becomes impossible to tractably analyze, as all the LLMs train off each other)

dTal··on Using A.I. to decode communication of sperm whales
I wonder what Wittgenstein would have to say about Alex the African Grey.
dTal··on Using A.I. to decode communication of sperm whales
Why inject such disingenuous chaff into the discussion?

Two options:

- You genuinely believe that all living organisms have equal moral weight. You are either a Jainist who sweeps the floor in front of your feet to avoid treading on an insect, or else a callous murderer who would dispatch an annoying human as easily as an annoying fly.

- You are a sane person who does not in fact believe that, and this comment was some sort of bad faith "gotcha".

dTal··on Mistral Large 4
Well my meta point was more that framing an apparent consensus as an "asymmetric conversation" seems pretty motivated.

But to respond to your point, most everywhere is great if you're rich enough. Saudi Arabia is pretty rad if you happen to be in House of Saud.

dTal··on Study: Claude, ChatGPT Offer Different Shopping Prices Based on Wealth
As predictable as water running downhill. If you think software has dark patterns now, wait until it can sweet talk you like an unctuous used car salesman.

First rule of digital sovereignty: all software you don't control will be used against you.

dTal··on Mistral Large 4
Proposition: Europeans have a better quality of life than Americans

Your observation: Online, Europeans assert they have a better quality of life than Americans, while Americans do not assert the reverse

Your conclusion: "idk maybe they're lying???"

dTal··on Mistral Large 4
Perhaps "generating wealth (for whom?) and technology (to what social effect?)" is not the be-all and end-all of quality-of-life metrics?
Page 1 of 34Next →