HNHacker News
TopNewBestAskShowJobs

blamestross

696 karma · joined October 16, 2016

Mad Computer Scientist https://miasma.space

distributed hash tables, p2p systems, supply chain and network security.

submissionscomments
blamestross··on My agent.md to improve LLM-assisted code quality
The most powerful change I have run into is: "Positive phrasing" as a default, prefer to tell the model what they should do, and why. Not a prohibition on a behavior.

When you say "don't do x" you are just pre-seeding the model with "x" and the prohibition mitigates that some, but not as much as never having put "x" in the context in the first place.

"Do Y, for these reasons" can be shaped to achieve what you mean by "Don't do X"

"Don't do X" leads to "Wait, I need to make sure I didn't X" and "Let's look up X to make sure I don't do that." And each time the odds of X happening keeps going up, not down.

blamestross··on White House Says Slavery Is Being Taught Wrong, Should Not Be So Negative
I just want it to be more commonly understood that enslavement is perfectly legal and enshrined in the 13th amendment. We have a million, mostly black, perfectly legal slaves in the US right now.
blamestross··on My Dad Is in Prison. Here's Why I Built an AI Chatbot for His Electronic Tablet
My experience is that the LLM will be less creative than the human in this situation. Necessity is the mother of invention. Prison is a special sort of neolithic age.
blamestross··on Private prisons announce $1.4B in revenue as immigration detentions climb
What does that look like? Dropped in a foreign country without money or documents. Likely in the custody of a government who doesn't care.

How do you re-enter the country?

blamestross··on Vint Cerf, “father of the Internet”, is retiring
My "When I met Vint" story is less exciting then some here. He attended the "Disability Support Group" at Google regularly. He made a point of just being there to listen and support others.
blamestross··on More evidence is consistent with possible ancient life on Mars (2025)
We have long passed at least my owner personal "threshold of evidence" to assume that Mars had life.

That bar is pretty low. I am not hard to convince.

We don't have a baseline by which to contextualize the situation statistically. There is no p-value we can meaningfully satisfy. Worse than that, evidence required for a belief derives from the risk of being wrong.

Is there a discussion of what evidence WOULD be enough for the scientific community to say loudly "There was life on mars". Because right now it seems all risk and no utility, so the evidence bar would be really high.

blamestross··on The unbearable cheapness of open weight models
Well I can't speak to the chinese gov part, but ALL the models are IP laundering systems. I'd rather IP get laundered into open source.
blamestross··on The anatomy of an AI-native org
This is all built on a premise that kinda scares me. Its not a new idea, it's one I have participated in my entire career. That concentration of power at the top labeled "why". It doesn't actually make sense, it just gives the people with power a narrative of security. "Why" should be a conversation up and down the org chart. If "what" doesn't come with why, you make the wrong "what". If "how" doesn't come with "why" you sabotage yourself. If your "why" can't survive effective communication and context, it was probably a bad call. If you feel like you "need to go fast" you are just gambling and looking at the world with survivorship bias.

Every engineering decision and debate I have seen has always boiled down to 2 things "people don't know the motivation for an action" and "people have incomplete knowledge of the situation". Normally both. Every single conversation, when poised in a "this is the high level goal" context, went from a debate to a constructive design session in seconds.

Failure to "propagate the why" is a tar-pit for execution and decision making.

blamestross··on A tale of two path separators
Someday ASCII 28 thru 31 will be loved.
blamestross··on Should everyone be on GLP-1s?
The basic concept, "put a new compound in everybody's bodies" because it makes them better suited for thier environment is entirely reasonable.

Humans didn't evolve in the context of plentiful food. It's one of the ways our success took us to a place our bodies are maladapted to. Genetic pressures will solve the problem eventually. Memetic measures have seen some success.

Like "flouride in the water", it is going to require a lot of work to justify that "exposing everyone is univerally good" but it is a possible reasonable and safe path eventually.

blamestross··on Superintelligence: The Idea That Eats Smart People (2016)
I think you are taking an entire space of "intellectual immune system" out of consideration. More than that, you are ignoring the core reason I think they are bottle-necked. They might have more access to compute. But to compute what? The bottleneck of intelligent behavior isn't compute it is experience. We have a lot of "text encoded experience" to feed it, through our collective corpus of writing, but ultimately potential behaviors can only be tested by active experimentation. No amount of observation can discern correlation from causality. Only active experimentation. The "train on itself" only works in a "toy universe" where the model of consequences are trivial to "test"

So in order to scale, AI doesn't need compute. It needs "engagement with reality and agency". Which is STILL might do better than us, but is happening in the real world, with real competition over resources. As long as we don't do something dumb like enthusiastically give it control over our major economic actors. I don't think we need to worry.

On the "intellectual immune system" side, I would argue that language's limitations are themselves fitness. We are already in danger of memetic hijacking. All those points you make about multiple instances of an AI cooperating, don't take malice and memetic attacks into account. It goes back to "why I am not afraid of grey goo". I trust yeast to find a way to metabolize basically everything. We have memetic attacks too.

blamestross··on Superintelligence: The Idea That Eats Smart People (2016)
As a distributed systems engineer, we are a LONG way from "magical scalable ai".

The bottleneck for a developing AI is experience. Yes we need compute, but we need data to compute on.

We have bypassed that limit by starting with literally every scrap of human generated prose that ever existed. I expect an explosion of expansion when visual and world models hit critical mass to properly leverage new experiences. But even then, engaging with reality is the bottleneck.

I can build you a very efficient scalable online map-reduce-like that runs inference on new corpus. We already made that. It took hardware getting large enough to fit the corpus in memory, instead of "scaling" it with networks for it to be viable. The latency of the network passing around partial solutions was WAY too high.

Computers don't scale forever. They are made of hot metals. The limits are heat, material, and the speed of light, but those are very real limits, that don't offer more than a constant multiplier of advantage over meat.

AIs might get smarter than us, arguably, like many other meat and paper based super-human intelligences around us, they already are. But it doesn't scale forever. It will hit limits, fairly quickly, of compute and experience to integrate into it's overfit model.

blamestross··on Superintelligence: The Idea That Eats Smart People (2016)
AI Superintelligence doesn't scare me for the same reasons "grey goo" doesn't scare me.

We are awash in self-replicating machines. The biosphere is already a grey-goo apocalypse. Any new competitors have a serious moat to cross to out compete any existing self-replicators.

We are awash in intelligent agents. Our society (and meta society) is full of superhuman agents already. There is a huge moat for any new intelligence paradigm to cross.

What I am afraid of is the existing superhuman agents (companies, governments and religons) will produce AGI or superintelligence and then proceed to use it as cognitive mitocondria, even further deepening thier supremacy in the cognitive ecosystem.

blamestross··on Unlawful by design: Exposing the human rights costs of generative AI
This is not how I wanted intellectual property to die. A special "IP Laundering" loophole for big tech companies is WORSE than the intellectual property issues we had before.
blamestross··on With Claude: Less Coding, More Testing
I found it kinda optimal to "incept" the llm. I better results if I just ask the right questions to guide to to a good solution.
blamestross··on Show HN: Freenet, a peer-to-peer platform for decentralized apps
Which isn't a web of trust. it is just an "allowlist". Humans are vulnerable to sybil attacks too.
blamestross··on Show HN: Freenet, a peer-to-peer platform for decentralized apps
Reputation systems have been a theoretical idea for a while, but we haven't come up with anything sybil-proof without centralized identity management. "we have a menu" sounds a lot like "we don't actually have any viable plan" in this case.

Don't get me wrong, this is awesome. I think it is built on a subtlety bad premise. I think it is time to start build organizational nomic games on this sort of contract system, literal organization governance, for systems like this to thrive.

blamestross··on Show HN: Files.md – Open-source alternative to Obsidian
Gods I love and loath Tiddlywiki. It has some of the most convoluted javascript written before javascript ever actually got all the features that made javascript convoluted. But it did the job!
blamestross··on It is time to give up the dualism introduced by the debate on consciousness
https://arxiv.org/pdf/1308.4375

I felt like this paper nailed it years a go, and nobody has followed up properly.

The metric involved is basically impossible to compute fully, but easy to approximate. Any online approximation will model everything it can see have changes until it is satisfied.

blamestross··on Security through obscurity is not bad
I feel like nobody actually understands the real reason security through obscurity is so bad. It results in dead cryptographers. The implementor becomes the weakest link in the chain, and entities with a tolerance for violence can fix that problem.
blamestross··on A more efficient implementation of Shor's algorithm
Seems like a wonderful use case for a one time pad! Every time you offer up the encryption key, it shows you were right!
blamestross··on Are We Idiocracy Yet?
Every time Idiocracy comes up, I feel obligated to point out that it is WILDLY optimistic. The people are dumb, not evil. They struggle to adapt and learn, but are willing to try and willing to accept new information with evidence.

We are not so lucky in reality.

blamestross··on Show HN: I Built Paul Graham's Intellectual Captcha Idea
Reminds me of IQ tests I took as a kid.

"Finish the sequence" with 4 options and "no pattern" as the choices.

It becomes "what does the moderately intelligent person who wrote the test thinks counts as a pattern" not the intended exercise at all. There was never enough samples to even guess at a real pattern in them.

blamestross··on Show HN: A P2P messenger with dual network modes (Fast and Tor)
Nobody does "face to face" key exchange like I imagine. Just two phones facing each other spamming QR codes for the other to read.

What I REALLY want is an app that builds a big bank of nonces between you and your peers over short range radio or QR codes and then lets you use a 1-time pad.

Ultimately, I'm only offering criticism because I have spent a lot of time working on exactly this problem, but I am not in a position to actually implement it. This is awesome and you should be proud of it.

blamestross··on Show HN: A P2P messenger with dual network modes (Fast and Tor)
So DHT robustness against censorship is superlinear of the number of participants.

The "break point" is when a DHT gets big enough I can't realistically MITM all the links with nodes "closer to the target" than existing ones.

This means big networks are great, small ones are cheap to just break. Its hard to skip the messy bootstrapping phase.

I'd encourage protocols to only rely on DHTs for small key-value stores if there isn't a trust mechanism in place to validate new peers.

Otherwise, all I have to do is mine for O(n^2) dht keys that cover the network. Figure out what your key mining difficulty is and you can identify what the cost would be.

blamestross··on Show HN: A P2P messenger with dual network modes (Fast and Tor)
"Mainline DHT" is the primary one. It backs magnet links and has resisted censorship for over a decade. Largest most robust group of cooperating computers there is.

You can reliably use it to store arbitrary key-value pairs up to 1kb in in size.

https://en.wikipedia.org/wiki/Mainline_DHT

blamestross··on Show HN: A P2P messenger with dual network modes (Fast and Tor)
I'm a fan of "face to face mutual qr code key exchange." I should implement that someday.
blamestross··on Show HN: A P2P messenger with dual network modes (Fast and Tor)
I don't recommend DHTs with public participants being made from scratch. Use mainline bittorrent DHT instead. Small networks are really easy to eclipse and censor.

DHTs of trusted participants are great.

blamestross··on Add age verification to accounts service
So every distro out there has a GOAL it exists for a REASON. If achieving that reason requires that compliance. Then comply.

Ageless Linux exists for a simple reason. It does not comply for the benefit of that reason.

Every other Linux distro has different goals. Thats why they exist.

blamestross··on Add age verification to accounts service
"I'll let the system admin set a birthday for the user" is exactly how Linux should comply with this legislation. That is what this does.

I disagree with this law. I like that https://agelesslinux.org/ is challenging it. Expecting the rest of Linux ecosystem to actively NOT comply with a law isn't the right response.

← PreviousPage 2 of 21Next →