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 All my thoughts after 40 hours in the Vision Pro
The problem is that it isn't him making the sacrifice, it's his family.
blamestross··on All my thoughts after 40 hours in the Vision Pro
https://149909199.v2.pressablecdn.com/wp-content/uploads/202...

This is the most ironic image I've seen in a while. "Even though I know all growth curves are sigmoid, I'm going to pretend the long term curve is exponential not also sigmoid"

blamestross··on Spider webs capture environmental DNA from terrestrial vertebrates
Spider Civilization.

I like TFA a lot, but it is very character driven. CoT is a collectivist story, the non-human-centric story reads like David Attenborough is narrating it. I like that part a lot.

The human/character driven parts of the story feel a bit tacked-on and not actually critical to the themes and messages of the story.

I think it is worth a read, and the sequels are more character driven than CoT.

blamestross··on Disney unveils the HoloTile floor
The normal model of friction already has static vs moving friction. They just vibrate to switch between them.
blamestross··on The Enchippening
I feel like a lot of people need to learn what a sigmoid curve is, and how only some of it "feels exponential".
blamestross··on "Elected leaders must tax us, the super rich. We'd be proud to pay more."
> Or we can stop trying to have the government solve every problem there is and stop taxing people at every instance of money transfer. If you’re an average person your income is taxed, your spending is taxed and any investments you make are then taxed. If you give someone a gift it’s taxed ffs. The end result is a government that spends more than it makes so think tanks can push the idea we need to find more ways and people to tax.

That literally is what I am proposing. Things like pollution, global warming, and resource extraction should have the costs on the commons paid back to the commons. If that cost is too high for profitability, the business shouldn't exist.

blamestross··on "Elected leaders must tax us, the super rich. We'd be proud to pay more."
Honestly, I think we need an entirely different way of looking at "taxes". We need to find a way to calculate the externalities of a company and charge that as taxes.

I don't have a magic way to do that, nor am I convinced it is possible.

But I think that is the bare minimum we can agree government and regulations are for, to prevent and compensate for us hurting each other. Especially in ways that are hard for individuals to handle.

blamestross··on Filecoin Foundation Successfully Deploys IPFS in Space
Well the caching aspect of IPFS is useful for space, the transport protocols are way too chatty and latency sensitive for real usage.
blamestross··on Psst: Paper-Based Secret Sharing Technique
Honestly I am more excited about the python code that generates the SVG workbook. I am going to borrow that technique!
blamestross··on Incompetent Middle Management Killed Google
Try looking at that ratio over time.
blamestross··on Incompetent Middle Management Killed Google
I feel a much stronger urge to blame the c-suite.

Ultimately it is so much simpler. Google is a growth stock. It's value isn't in revenue. It isn't even in the YoY growth. The value is in the second derivative of revenue. It's value is predicated on the company staying on the "fun and super linear growth" part of the sigmoid growth curve.

The growth is changing inflection and they have to do everything they can to keep the illusion that it isn't alive because once Google stops accelerating the stock will crash and it will "fail".

So the c-suite could manage the transition from a growth stock to a dividend stock, or they could keep chopping off bits to keep the old dream alive for one more quarter.

blamestross··on Thousands of AI Authors on the Future of AI
Profit motive
blamestross··on Designing, manufacturing, and selling an LED 'social battery' pin badge
Totally the same on distracting myself from pain by talking to others pattern! It is a big part of why I put myself in social situations despite being in pain.

At the end of the day, it is all crappy 1st order approximations of ideal behavior. That is the best humans can do. So we should try and make sure the "error" on those gets spread around/rotated in time rather than consistently excluding anybody.

blamestross··on Designing, manufacturing, and selling an LED 'social battery' pin badge
I have multiple chronic pain and health issues that can make extended social situations hard for me. Often it sneaks up on me.

"low social battery"/"spoons" doesn't mean you should avoid them or try and fix anything. What you should do is take it as a signal to not take offense and be patient if they need to leave suddenly, get distracted, or have difficulty. You can be welcoming while still making them some space to be tired or in pain.

My experience is that people who are lacking "social skills" aren't actually lacking social skills, they are lacking "social instincts". For people who have them, those social instincts work great 80% of the time, and most people just let them fail 20% of the time instead of applying critical thought to the situation (taking offense where none is meant or just miscommunicating).

Bluntly put, this tool is likely not meant for your consumption and your "lack of social skills" probably makes it much less necessary. You have to use critical thought every time you are in social interaction, so a reminder to do so is just signal noise to you.

blamestross··on O(1) Build File
Don't worry, pretending an `O(lg(n))` thing is `O(1)` thing never misleads or causes problems.

Actually it doesn't really, but that is a cultural time bomb in our thinking that will catch up to us someday.

blamestross··on America lost the chestnut, its "perfect tree"
And still have a problem with having too much bamboo growing outside...
blamestross··on Borges and AI
No it doesn't contradict what you explicitly said. It is intended to contradict whatever motivation you had to post a comment that wasn't actually productive. Data is valuable without needing attached argument.
blamestross··on Cory Doctorow: What Kind of Bubble Is AI?
The sort of deception that goes into making a bubble generally requires a kernel of truth to build and exaggerate on. Bubbles have always served to get investors to pay for training new labor for the industry. That depresses wages and labor costs in the long term.

The fact it isn't also triggering a hiring boom stands out to me as worth looking at.

blamestross··on Cory Doctorow Blasts AI as a Fraud-Filled Bubble
Yes, they do have use. They are also very oversold and literally being sold for jobs they cannot do. It does help write things that were either formulaic or entirely without real content.

Allowing them to do analysis or summary of anything that isn't already in a Wikipedia entry is radically dangerous because of how good they are at hallucinating and lying (and we can't detect that at scale). It is like an automated "yes man". This is the most dangerous bubble in living memory and the cost of it popping will be in the failure of any industry that is foolish enough to rely on them at scale.

And to attack that study in particular, "help people who couldn't write well write better" is complicated, increasing accessibility is good, but there is a BIG difference in an increase in writing quality versus an increase in productivity AND correctness.

blamestross··on Cory Doctorow Blasts AI as a Fraud-Filled Bubble
It feels surreal seeing so many comments like this here and in public discourse in general. The advertising/hype machine convincing people a low quality continuation model is the next coming of AI must be amazing.

LLMs have lots of cool advancements involved in them but isn't what you are being sold. Anybody who knows how these models work and isn't making money off them can explain just how limited they are and how they can't safely be applied at scale. Yet they are doing so anyways. Invest and lie while you can, and get out before it all falls apart.

blamestross··on Do large language models need all those layers?
The answer is always "No it doesn't need all those nodes", but we do it that way because it makes doing the math easier. I bet they just removed a bunch of edge weights to nodes, but used the same matrices for calculating output.

I wish arbitrary topology networks were scalable (I love NEAT) but bipartide graphs crunch good in GPUs.

blamestross··on Researchers create humanitarian law digital emblem for the Red Cross
They discovered public key certificate infrastructure and are calling it an "Emblem"
blamestross··on Cats that fetch are an evolutionary mystery
The trick here, is that once you got him his own seat at the table, he was perfectly happy to sit on it with his face poking up and watching us play. If I just made him get off the table, the first time we look away he would hop right back on and start knocking pieces off in protest of this mistreatment.
blamestross··on Cats that fetch are an evolutionary mystery
It sounds more like 15% of cats realized their humans could be made to entertain them with proper prompting.

My experience with fetching cats is that it is a comforting/boredom relief behavior. They bring me a toy, I throw and they chase it, then bring it back to my general location (they get distracted and drop it easily, most often right before jumping onto the couch). The fact the loop can be repeated a few times before they get bored again and move on is just bonus points.

blamestross··on Cats that fetch are an evolutionary mystery
I had a cat who loved to be tossed across the room (onto the nice safe soft bed) and who would excitedly run back to be tossed again. Just don't ask how we figured it out in the first place...

But he was wierd. He liked going on walks on a leash, and demanded a chair at the table when we played boardgames and he would happily sit and watch us play (or sit in the middle of the board until we got him a seat).

The first time I met him, he very clearly, intentionally guided me to his brother who was sick (meow, runs towards me, meow then back towards his brother, repeat).

We lost him kinda without warning, he got an ear infection that popped his eardrum then he never recovered from the shock. Cats don't communicate pain or discomfort well, so we didn't even realize something was wrong until it was too late.

But yeah, consistent recessive outliers are an evolutionary advantage trait. I've seen a lot of arguments that neurodiversity is an evolutionary trait, as having multiple behavior patterns in the population helps with survival. Cats are social animals, which means a similar neurodiversity pressure probably acts on them.

blamestross··on After OpenAI's blowup, it seems pretty clear that 'AI safety' isn't a real thing
AI safety is a glorified proxy for runaway late stage capitalism. Large Cooperations aren't really distinguishable from superhuman AGI. It is just more socially acceptable to blame AI for the problems and ignore the fact we already face them all.
blamestross··on OpenAI staff threaten to quit unless board resigns
It's like the "Open" in OpenAi was always an open and obvious lie and everybody except the nonprofit oriented folks on the board knew that. Everybody but them is here to make money and only used the nonprofit as a temporary vehicle for credibility and investment that has just been shed like a cicada shell.
blamestross··on The Changing "Guarantees" Given by Python's Global Interpreter Lock
Most people and scripts using the GIL to ensure safety likely don't even realize it is necessary. That is the hard part of this type of migration. Decades of code written with the implicit assumption of the GIL for all situations.
blamestross··on Ask HN: What Is Your Take on Stephen Wolfram's "A New Kind of Science"?
I think the fundamental idea of "start thinking in graph structures for spacetime instead of rubber sheets" is good and going to eventually pay off. But right now it doesn't offer anything testable and thus is useless for anything but thinking about while high.
blamestross··on A theory that Mars lost its magnetic field and then its oceans (2022)
Progress is exponential until "surprise it is a sigmoid!" I'd argue we are there already, but even if we aren't planting seeds is rarely a waste in the long term.
← PreviousPage 7 of 21Next →