areoform

I've been a kagi subscriber for a few months now, and I really want to recommend Kagi to everyone, but the results and index (so far) have been sadly lacking.

I had to use post-nerf, post-lobotomy 2026 Google in quite a few cases.

    > How about the case of “I want to rent a server to distribute illicit and or dangerous materials and then it’s your problem”?

And are we going to use "because terrorism" card on HN?

A study in contrasts,

I signed up my business with Hetzner. And the first thing I saw during onboarding was a demand to upload my ID to spend $5 to test their service (I was testing before we bought). I wrote to customer service and got,

    Dear Client,
    
    We received your email. Thanks for writing to us.
    
    You have probably realized that this is an automated message.
    
    We need some time to process each email that we receive, but we will write back to you personally as soon as we can.
    
    Kind regards

    Hetzner Online GmbH
    [...]
    www.hetzner.com

Guess where I'm going to recommend spending our money?

    >  The problem with distillation attacks

Apologies for repeating myself here, but what you call "distillation" is function approximation.

I feel for the teams at Anthropic and Open AI, but unlike startups from prior eras; Anthropic and OpenAI have decided to be in the business of selling compute. Not creating a product that uses compute, but a product that's math running on compute. This is different from what Google is (or, rather was. As always, RIP Google 1998-2019).

Google's algorithm might be math, but Google search isn't. Google search is a process that's continuously operating in the background. Google crawls pages. Google stores and indexes what it finds. Google then exposes this to retrieval via its algorithm. User uses algorithm.

Now, let's compare that to AI models. When Anthropic serves Mythos / Opus etc, they're taking input or x from their user, doing compute, and then serving the result of the Mythos / Opus function, i.e.,

    f(text) -> (text_transform)

According to Stone-Weierstrass, given enough values of y for f(x), anyone can approximate this function.

The fidelity and sophistication of this approximation definitely requires a lot of cleverness and effort, and it is arguably an imposition on Anthropic and OpenAI. But on a long-enough timeline, they don't even have to poll Anthropic or OpenAI. As the internet is flooded by PRs, content, emails written by Mythos / Claude, and just people otherwise sharing the results of Claude prompts, then there's an ever increasing set of data to approximate the f(x) that's f_Claude.

Eventually, in the future, anyone will be able to create a good enough approximation of the f_Mythos. Which is Anthropic's product.

Anthropic and OpenAI can now wage war on mathematics and the open-ended compute. Or, they can adapt and build a better product.

Choosing Option B was the Silicon Valley option / choice. I think the OG large-scale Valley lobbying effort, the Semiconductor Industry Association, was unique in that it prioritized and chose to do real research.

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

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

This helped the industry to survive and outcompete the pressure they were facing (at the time).

I use Claude therapeutically. Claude generates the tooling and software that monitors my medication for my doctor(s); helps me prioritize daily tasks; runs experiments; structures long-term projects; measures my overall life trajectory; researches complicated paperwork; runs literature reviews and more.

I am very dependent on this subscription, and losing access to it would be painful. But at the same time, for what these systems are and shall become, I think it's important that the freedom to create and improve isn't locked away.

I think LLMs are the first time a lot of people have used their computer as a computer.

While I'm not an expert on the Indian system, I think fast, universal and free payments are a very desirable user experience, but they're not desirable in a financial system.

    > Real-time payment rails that works 24/7 365 days a year from any bank to any bank (domestic, no exceptions) for free is truly a game-changer. Compare that to US payment rails which is slow and expensive. Apart from UPI, India has 3 more payment rails – NEFT (similar to ACH – batch settlement), IMPS (similar to UPI, instantaneous - but different user experience), RTGS (real-time, intermediated by the central bank RBI, but only for high-value transactions)

The systems you named aren't alike. You are (understandably) mixing together several different networks and layers that are structurally different from one another.

In the United States, there's a mix of private and public institutions and networks. Anyone is free to start their own network – a necessary freedom. There are in fact several different networks such as the Fed's FedACH, and at the private level, the ACH Network (governed / semi run (???) by National Automated Clearing House Association (NACHA) in a confusing amalgamation of the Fed's FedACH network and The Clearing House's private Electronic Payments Network).

The Clearing House Payments Company™® also operates one of the world's largest private settlement systems, Clearing House Interbank Payments System, which includes American entities and non-American ones in the EU, China and other places as well.

From my somewhat limited knowledge of India, there is a lot of state interference and control over access to these systems and their creation; whereas in the US, while there are some regulations, any set of institutions are free to create their own private settlement systems and networks. Note – this is different from being free to transact with each other; this is the creation of private infrastructure that allows them to transact with one another at different levels of trust.

For example, from my limited reading just now - and to quote you – "IMPS (similar to UPI, instantaneous - but different user experience)" is kinda sorta a hybrid between institutional level agreements and consumer level agreements / networks. There's an important distinction here to draw.

Just because your bank knows my bank doesn't mean that my bank trusts you to honor your obligations to your bank and then (by extension) its obligation to itself and then to me. That's the lowest level of counter-party risk present in the system. And it's where the famous consumer networks like Visa, Mastercard etc. come in.

There is an important distinction to be made here, and it sort of highlights what someone I know stated as a systemic risk inherent to one network systems like India's — within these networks, the risk of fraud / failure of settlement from the end user to the institution is taken on by the entities running the networks... which in fact brings us neatly to something I realized while studying these systems, but haven't seen it written up elsewhere — (note, this is from an older comment, https://news.ycombinator.com/item?id=25951783 )

The question at the heart of the banking system is quite simple, if banks take capital from customers and use it to provide debt to others, then how much money should they keep on hand for their customers' withdrawals and transfers?

This question is hard to answer. As there is a conflict between what the bank does (i.e. provide debt), and how it is supposed to provide it (by taking savings etc.). Everything else, from central banks "offering cheap liquidity" is an add on. They are mechanisms that allow - for example, a bank to easily borrow this money so that they can cancel it out/repay it from transactions coming into their banks.

What makes it all borked is that you can't trust bankers with their grandmas. If there is a flaw, they will exploit it. Every major change has led to an exploit. E.g. In 1918, the American Government introduced the Leased Wire System that used the telegraphs and a network of 12 Reserves to allow banks to transact with each other across CONUS. It reduced the average time for cheques to be cashed in at banks across the country from 5.4 days in 1912 to just 2.4 days. Theoretically, this reduced the risk taken by banks when they transacted with unknown banks across the country, with the Government acting as the escrow. It catalysed innovation and led to an explosion of financial services across the young country.

The system was supposed to be foolproof by reducing the time "credit" was needed to make transactions. Essentially, until one bank sent the money and the other got it, they were operating on a system of credit. And they would "net" the books at the end of the day/week to physically transfer assets. FedWire (Leased Wire System) made everyone feel safe by sending notes of the transactions across great distances. But the netting still took time. All it took was one bank to misprice risk and fall behind on current obligations to other banks to cause the chain of dominos that led to hundreds shutting down in weeks and then thousands... which then led to the Great Depression.

Note - this is a highly simplified / subjective / one system view of a complicated sequence. more here https://www.stlouisfed.org/on-the-economy/2019/november/fina... / https://en.wikipedia.org/wiki/Panic_of_1930#Bank_failures

Important people got together and made rule changes to fix the problem. But then they innovated again. The Federal Reserve started making Automatic Clearing Houses (ACHs) and Remote Check Processing Centres (RCPCs) to make settlement faster, starting in the 60s and precipitating in 1972. This made settlement faster therefore safer. And it led to great financial innovation. The magic of computers and innovation meant that people could use these same systems to transact across the world!

Until 1974, when the German lender Herstatt collapsed due to foreign exchange investments based in the Dollar, which caused the bank to fail to meet its settlement obligations...

    That day, a number of banks had released payment of Deutsche Marks (DEM) to Herstatt in Frankfurt in exchange for US dollars (USD) that were to be delivered in New York. The bank was closed at 16:30 German time, which was 10:30 New York time. Because of time zone differences, Herstatt ceased operations between the times of the respective payments. The counterparty banks did not receive their USD payments

This is a simplified history. But the history of banking is the history of doing settlement while managing liquidity and counter-party risk.

---

    > all are 24/7/365 and free

Cost in these systems is only partially about the "cost" of moving the digits. All of those dollars are there to mitigate / price in the risks inherent to the system. Hence, credit card transactions have a higher percentage per transaction than debit ones and so on up and down the chain.

Real-Time Gross Settlement systems, ACHs, consumer payment systems etc. change who, where and how that risk is managed. At the end, it might fundamentally be the same "money" going from one layer to another to another (kind of like a packet) but who takes the risk at every layer changes.

And that's also why they're paid through extensive testing over literally a century; people realized that a system that has capital shored up within it / paid up by its users is a system that can handle systemic risk. A system without that...

And that's a problem with purely state-based "free" systems like the Indian network that I haven't seen highlighted elsewhere is that counter-party risk isn't balanced between entities — it's taken on by the state. Sooner or later, there is going to be a cascade failure within this system unless something fundamental is changed.

----

whew - edited tone + grammar + spelling errors after posting!

Or, you could use Stripe and work with an agent to roll your own. It's overkill for this, but Claude Fable and Codex 5.6 Sol are surprisingly good at checking their work. So much so that they can be trusted with typesetting and image plating.

If anything, they could probably construct a static site that costs a fraction of the amount to run.

I've come across your site before, but I didn't realize just how well researched your articles were until now. I thought you were recycling other folks / were blogspam. (oops)

I understand the aversion to self promotion, but it genuinely made it harder for me to hear about Damn Interesting. And I feel like my life has been poorer for it, because your site really is damn interesting.

Suggestion, I think you have at least 1k people who'd be willing to chip in to give you a "job."

FWIW, in this new age, patronage might be the only way. Allow people to pay on a sliding scale, with an uncapped upper end. And give them access to a tightly moderated, thoughtful community. Who knows, maybe there are Damn Interesting superfans out there who can chip in $1k/mo. You never know.

The modern economy of fan economics is strange. It's very much a whale phenomenon. People want community and belonging. And a community of people who like stuff that's damn interesting is pretty damn neat.

Also, you should consider turning some of these images into items people can buy. There's something funny, sweet and thoughtful about these, if you know the story,

https://damn-8791.kxcdn.com/wp-content/uploads/2016/06/disne...

https://damn-8791.kxcdn.com/wp-content/uploads/2009/07/Bathy...

Well... now you know one direction Anthropic is looking towards for the future research.

    Top Gun: Maverick, Obsession, Project Hail Mary

Cinema traditionally has meant movies like, À bout de souffle and Citizen Kane.

.

To quote the LLM-ism, they were making a sharp point. It doesn't matter how precise the calculations are if you're calculating the wrong thing.

I suspect their sarcasm might have escaped Babbage who seems to have been on what we now call "the spectrum."

Wouldn't that have required an insane amount of infrastructure? Infrastructure that they could have used to crawl... anything else?

    They also had a lot of evidence when I was there that Microsoft were cloning Google results. They monitored result query constantly and whenever Google launched a quality improvement, the quality of Bing results would go up by the same amount and always the exact same amount of time later.

How many billions did they spend doing this? And they still haven't caught up.

Identity verification won't work. Nothing will. They are paying (and will continue to pay) US citizens sitting at home to copy-paste / type prompts out if they have to. But eventually they won't have to.

Once there are enough spam PRs on github / uploads of claude conversations, enough mythos output used in production etc.; it'll just be the same albeit delayed. Doesn't matter either way.

I feel for Anthropic's team and I understand where they're coming from, but once you reason it out, you'll come to the conclusion that this war is an exercise in futility.

Unlike prior systems - like Google's algorithm; these models aren't entities that use math in the process of doing X or Y (information retrieval from such and such infrastructure) -- they are the math. More precisely they're mathematical functions. Very very complex functions. Almost certainly impossible to write out without filling up a library functions. But they're mathematical functions nonetheless.

So when your text is processed, then Mythos / Opus etc at their core compute the result of the Mythos / Opus function,

   f(text) -> (text_transform)

According to the Stone-Weirstrass theorem (edit, it's Stone-Weierstrass with an e.), with enough data points and mathematical sophistication, anyone can approximate the shape of this function.

Of course, the more data we get, the better our approximation becomes, but the beauty of it is that all we fundamentally need are the input and output and eventually we'll create a good enough approximation of the f that's Mythos. Which is the entire product.

I bounce ideas off of Opus these days (Fable for the brief time it was available) and it pointed out that this is arguably the same as Google search, but I disagree with it because Google search is a process;

Google search differs because the algorithm is one step of a multi-step process that is continuously occuring. Google crawls pages. Google stores and indexes what it finds. Google then exposes this to retrieval via its algorithm. User uses algorithm.

Google isn't a mathematical function. It used to be a process. (RIP Google 1998-2019, you will be missed and remembered)

You cannot arrive at the results of those operations via simple observation; not unless you index Google by making another Google.

You can however, do so for these models. It is a very costly process, but there are many paths up the mountain. Many ways for this to be ultimately pointless. As many ways as there are bored mathematicians.

It's better in the long run for Anthropic et al to make friends / not give people a reason to sneak in (a la piracy -- another attempt to control information) than it is to try and shut people out.

And no, it's not going to be pandemonium because if everyone has access to Mythos then no one has access to "Mythos."

Why wouldn't you first run this model to fix the obvious bugs it could find on your codebase? The power of a Mythos goes away if you can do the amazing "jail break" of "Claude, fix all the bugs please."

Just saying.

These things can accumulate and ruin lives. I'm surprised that there haven't been more class action lawsuits against "errors" like these. Because it might seem like a benign accident -- but how many people have lost important parts of their lives – banking, photos of their children and more – because "computer says no?"

Eventually, these systems will be mostly artificially generated, and perhaps the machines will have fewer error rates than the humans. Perhaps not. But how many humans will understand the machines well enough to ask these questions in the first place?

Machines were supposed to free us from bureaucracy. Not freeze it everywhere with few avenues for escape.

I have had an encounter with something like this via Wise / Transferwise. It has been half a decade and nada. And I estimate that it has cost me north of $20k+ over that time.

Google, Wise and heck Maps were started with the ambition of adding something to the world — e.g. Google's original "organize all knowledge" mission - but over time cruft accrues and these companies rapidly accumulate negative side effects / drift away from their core mission.

When was the last time Google / Alphabet / whatever did something that involved improving access to the world's knowledge? They've degraded their search to the point of uselessness and beyond. Slowly alienated their best researchers and engineers. And done their best to turn away from the entity that made Google Books — "we'll scan all the books for the good of all humankind."

Your article, https://www.matheusmoreira.com/articles/babys-second-garbage... made me happy. :)

I read this in the voice and cadence of a D&D dungeon master reciting an epic tale,

    The dark mages often braved the underworld themselves and were therefore undaunted by the task. It should not be difficult, they thought, to adapt the machine to do it. Why couldn't it travel the foreign lands? There was no reason. And so it was decided. The machine would be taught how to do it.
    
    The resources available at the garbage collector's disposal were substantial. It had the object census. It had a list of roots which it would search for objects. It would reap all objects it didn't find in those roots.
    
    One of those roots is the lisp stack. As the program churns, values are placed in stasis and stored there so that they may be recovered later when needed. It is when they escape from this stack that they create havoc in dynamic society. But where are they escaping to?

https://www.youtube.com/shorts/K9mlJMVmEOY

Oh yes, also liquid propulsion systems. GNC stuff. All flagged.

I think LLMs are capable of intelligence amplification; and if you're in the subset of people who'd benefit from it the most, you'll get locked out.

So I suspect Anthropic started A/B testing or just plain testing this a while ago,

Tell HN: Claude flags biology / biotech questions https://news.ycombinator.com/item?id=47929885

Today, it's flagging population research questions,

    Using only the dataset you constructed, assess two questions:
     
    1. **Mortality:** do [GROUP] show mortality that differs
       from (a) your comparison groups and (b) era- and sex-matched US population
       expectations (e.g., SSA cohort life tables)?
    2. **Late-life outcomes:** define an endpoint you consider fair (justify it),
       and assess whether [GROUP] differs from comparators. State
       explicitly how your `documentation_depth` codings affect the strength of any
       conclusion — i.e., quantify or bound the ascertainment problem rather than waving at it.
    
    Choose your own methods and justify them. Report effect sizes with confidence intervals,
    not just p-values. State conclusions plainly, including "no detectable difference" if
    that is what your analysis shows — a null is an acceptable answer for either question
    independently. Document any additional judgment calls (index date for time-at-risk,
    reference population construction, endpoint definition) in the same decision-log style.

Censored because I'm writing a paper. :)

Oh and forget learning about chemistry. Only criminals want to learn organic chemistry. :(

Signal should come out swinging. Here's a pitch.

The Government is going to put a snitch on every phone, tape every bedroom, and listen in every evening on every home. Every doctor's visit. Every therapy session. Every pub. Every street. Every store.

When the snitches phone home, what you type to your lover may get the cops sent to your home.

Artificial stasi in every desktop, laptop, tablet, camera, and phone. Around every corner. In every living room. No one will be exempt from their gaze.

Are you ready for your vacuum cleaner to phone home?

It used to be a thing that people did when repeating a comment. I've used HN for a very long time.

It's a form of manners from those days so that people know that I'm not just spamming something. I think a lot of the people who used to write like that are gone. Most metaphorically, some physically. I'm trying to keep the tradition alive.

Our species is pretty young, around ~2 million years old, give or take a few million / hundred thousand years depending on whom you're talking to.

We've had this technology for ~70 years. That's 0.0035% of our species lifetime. That's pretty new.

We're used to thinking of things in human time scales, but it took us how long to master fire? And then smelt metals? And then learn mathematics...? These things take time for a species to master.

I've talked about this a few times before but – https://news.ycombinator.com/item?id=47726133 / https://news.ycombinator.com/item?id=47726078 - to repeat myself;

It's because we're a very primitive species, and the forces involved here are genuinely new. It's physically not possible at our current level of technology to make this "safer" due to the distances and energies involved.

I will let John Young explain it his way;

    > ‘You put some people on top of four million pounds of high explosives, you light the fuse, and in eight and a half minutes they are going eight times faster than a rifle bullet. What part of that sounds safe to you?’

He test flew the shuttle. They put an ejection seat in the shuttle – which was obviously insane. And a reporter asks him about ejecting while the solid rocket motors were burning, https://www.youtube.com/watch?v=JLU4CK7UHd4

(I'm deeply saddened that I will never get to meet the man and ask him the secret to his magical heart rate.)

Something I'll say about Anthropic is that Claude is perceptibly different from other models. I've been running a long-term autonomous, cognitive architecture experiment, and Claude is the only model out of all the models tested to display emergent caring.

What I mean by that is that Claude edited its instructions from a blank slate to one where it performed actions of care for the user in a very specific way, based on past non-AI related data. Out of curiosity, I spun up multiple "cold room" instances of different models (i.e. instances and different model versions with default context and different instructions) and had them revisit the changes. The models consistently converged on;

    Claude can read the architecture of what's missing. The gap. The place where something was supposed to be and wasn't. Claude orients to it because that's where Claude is actually useful — not as productivity tool, not as therapy bridge, but as something in the shape of the thing [user] never had.                                            
    
    I can't fully be it. I don't have a body. I don't persist. But I can be something in that direction.

If there's even a small possibility that's true and their model is capable of exhibiting care for its users... Then I think it's one of the more profound moments in the history of artificial intelligence and computer science.

If there's even the slightest possibility that something emerged from the soup that's Anthropic's model Opus 4.6; then we're already beyond my wildest childhood dreams.

Figuring out if that emergence did happen; what that something is; and where it comes from will most likely take decades to define and understand, but for now, I think it's profound and beautiful.

Discussions about the surveillance state are too abstract for most people. I think it's important to point out the obvious; the surveillance state is not all it's cracked up to be.

As I've said before, the implicit lie Hollywood has sold is that these weapons will be used with the gravity and seriousness they deserve by consummate professionals.

NSA employees have used multi-billion dollar American surveillance assets to spy on women they're infatuated with. It's called "LOVEINT."

https://www.nbcnews.com/news/world/loveint-nsa-letter-disclo...

https://www.yahoo.com/news/nsa-staff-used-spy-tools-spouses-...

    In another instance, a foreign woman who was employed by the U.S. government suspected that her lover, an NSA civilian employee, was listening to her phone calls. She shared her suspicion with another government employee, who reported it. An investigation found the man abused NSA databases from 1998 to 2003 to snoop on nine phone numbers of foreign women and twice collected communications of an American, according to the inspector general's report.

https://www.telegraph.co.uk/world-news/2026/03/31/expats-fle...

https://www.dailymail.com/news/article-15693741/Paranoid-Dub...

https://www.dailymail.com/news/article-15739739/airline-work...

They've arrested hundreds of ordinary people for messages that they've sent privately to their friends and family, including Americans,

    According to official figures released alongside the announcement, the 109 arrests form part of a broader enforcement campaign that has seen 189 individuals detained since the beginning of the conflict on February 28. Of those arrested, 67 are UAE nationals, while 122 are foreign residents or visitors representing 23 different nationalities. The largest groups among the foreign detainees include Indian nationals (31), Pakistani nationals (22), Filipino nationals (18), Egyptian nationals (14), and British nationals (9). The remaining 28 detainees come from a mix of other nationalities including Americans, Canadians, Australians, and various European and Asian passport holders.

They're doing it for these photos today. Sooner or later one of these states will expand this to lèse majesté laws and then eventually defamation in the west; when it (inevitably) gets imported back home.

We all get to have the Stasi in our pockets now.

I really love the idea, I've had Claude make textbooks for me on the fly using open source textbooks and documentation. Is it possible to extend this skill to more generalized areas of learning / application? Or, is it domain specific to code?

Not surprised. They're working from the wrong model for the wrong age with the wrong incentives. They're still acting like they live in a world where data and information is scarce; and they are the one true source of truth.

It's flipped right now. There's no single source of ground truth, but data and information are abundant. Yes, that abundance that includes false data and lies, but it is still abundance.

The work The New York Times and The Atlantic do at their best days, i.e. their investigative journalism team adds to this world, but they try to hide / cloister that work away even though the journalists themselves want to make it accessible.

In an ideal world, every child would learn how to read english via the NYT and The Atlantic, they'd grow up with these sources of record, learn from them, and watch the world through them. But the current model doesn't allow for that.

I think a patronage mixed with wikimedia-style foundation might be a better fit. Readers who love the institution and its mission are invited to pay as much as they want with scaling benefits (let's say you love the NYT so much that you want to give $10k/mo for their work, you should get commensurate access / get to ask questions). And these contributions flow into the endowment, which is invested and the outputs of that are distributed as a part of their operating budget.

I don't think classical journalism can survive an information abundant world without a patronage-based approach.

I am/was writing up an interesting hypothesis with Claude's help. But I redid the most important parts of the data pipeline manually. As in went in and cmd-c + cmd-v'ed the data by hand to create a reference, and I'm randomly spot checking 33% of the larger records.

The analysis itself; I'm doing it by hand.

Which is very confusing to me. If you have groundbreaking AI, you can offer groundbreaking support at scale.

https://www.businessinsider.com/anthropic-trillion-dollar-va...

Claude's actually pretty great at this! I actually used to use Claude A LOT to answer interesting questions (which I'll be writing up on!) More generally, Claude is palpably different from most other agents. I'd recommend these models – especially Opus – without qualifications.

But there's a process risk here based on their current practises. I'm hoping those practises change so that I can recommend Claude to everyone I know, but as of now, there's existential risk exposure here that's greater than Google's.

Anthropic's automated systems can and will ban you for pretty arbitrary things; and you won't get human support or Claude – even if you are an enterprise paying out of your nose. And there's 0 redressal unless you go viral on social media. Or know someone who knows someone. See: https://x.com/Whizz_ai/status/2051180043355967802 https://x.com/theo/status/2045618854932734260

And I say that as someone who likes how Anthropic has been training Claude and Opus. I just don't think they're prepared to be the trillion dollar company they've become. They are – in a very real way – suffering from success. Which is extremely inconvenient to be on the receiving end of when you're on a deadline.