HNHacker News
TopNewBestAskShowJobs

A_D_E_P_T

9,453 karma · joined June 8, 2022

submissionscomments
A_D_E_P_T··on As A.I. makes law firms more efficient, clients ask: 'Where's my discount?'
There's a solution that most are overlooking.

Legal work in the past involved hiring specialist firms, at exorbitant rates, for specialist work. (And virtually all legal work is specialized.)

Legal work in the future might involve hiring one or two in-house lawyers to handle all work, at a standard salary rate. The crucial skills are simply editing and research. Current/near-future LLMs surpass virtually all specialists in domain knowledge and they do it quickly and for the cost of a few cheap subscriptions.

You only really need a lawyer because a business can't represent itself in court without one.

A_D_E_P_T··on The Truth Physics Can No Longer Ignore
People don't like it, but the "weirdness" of life, as it relates to physics, just completely evaporates if the universe is superdeterministic.

And that's actually overkill. Plain vanilla determinism would do it. As would a universal-wavefunction/no-collapse conception of quantum mechanics. (i.e. the universe has its own wave function.)

Point is, what the author decries does not require New Physics. It should suffice to re-evaluate different interpretations of existing physics. These matters are, after all, by no means settled.

A_D_E_P_T··on Ask HN: Is Opus 5.5 another step change?
Yes, Opus 5.5 and Astra-6 are big improvements over Opus 5 and Sol-5.6. I'm at a point where I can automate much of my workflow, even for complicated tasks, which I could never do before.

Also the new Xiaomi model, Mimo v2.6, is outstanding as a light and cheap model. For coding, the new meta is using Astra-6-Max for orchestration and architecture, and Mimo for grunt work.

A_D_E_P_T··on [dead]
Whether or not it's genocide is debatable -- a matter of semantics. That it's "the opposite of genocide" is a laughable claim and demonstrates how detached from reality Israeli leadership is.

It's obviously ethnic cleansing -- this is the Israeli government's stated intent, and it couldn't possibly be any clearer.

And it's obviously immoral to wage war on an unfree and captive civilian population. This is the province of history's worst villains.

A_D_E_P_T··on Mercury Is Shrinking, and Faster Than We Thought
> Previous estimates suggested that Mercury had shrunk by just a few miles since its formation. Now, scientists say it is shrinking up to 30 percent faster than expected. It might have even shrunk by as much as 14.5 miles, which is nearly one percent of its current radius of 1,516 miles.

Less than 1% over 4.6 billion years, at that proximity to the sun, ain't that bad. The sun is expected to swallow Mercury in about 5 billion years, so it's going to meet a violent end long before it evaporates away into nothingness.

A_D_E_P_T··on MiMo v2.6
I must second this.

I'm in Europe, and here the options for home appliances are usually German (e.g. Philips), Balkan (e.g. Gorenje), or Xiaomi. Xiaomi is the best by far, and it's honestly not even close.

Their home appliances are so rock-solid that they actually still surprise me. For example: I've gone from having to replace electric water kettles every six months to buying one from Xiaomi and never replacing it. (Nigh on three years now.)

I really have a very positive impression of them.

A_D_E_P_T··on Deodands put a price on objects that caused death
The article belabors the point a little bit. A deodand was any inanimate object directly linked to a human fatality -- from a runaway cart to a falling branch -- and it was usually condemned and destroyed. (Or surrendered to the King and then sold, depending.)

There was often some hair-splitting involved, but the law of deodands was on the books for hundreds of years in England, and it reflects still older Greek and Hebrew laws where inanimate objects that killed people were cast out or destroyed.

The railways did contribute to the end of the custom. But this is not simply because rail equipment was too expensive to seize or condemn. Richard Posner, in The Problems of Jurisprudence, notes:

> [The law of deodands] eventually became dysfunctional but, as Holmes explained, the old law turned out to have a new use. The victim of a maritime accident may encounter great difficulty in trying to obtain legal redress directly against the owner of the ship that did the harm, because the owner may live far away, quite possibly in a foreign jurisdiction. By analogy to the law of deodands, why not treat the ship itself as the wrongdoer, and hence let the victim sue the ship in the first port in which it berths after the accident? Until the suit is resolved, the ship will not be allowed to leave the port unless the ship's owner or master posts a bond. In this way the victim can obtain and collect a judgment without traveling to the owner's domicile. Admiralty law adopted this approach and still follows it.

The interesting question is: Why were trains supposedly fatal to the law of deodands whereas ships were not?

I think that it has much more to do with the decline of superstition from roughly the second half of the 18th century on -- courts began to view the old law as an irrational relic, and it gradually fell out of legal fashion.

A_D_E_P_T··on AI-generated posters don’t have to be horrible
What's interesting is that you can't do the same with long-form AI prose. You can ask any given LLM to write in different styles, but it will always fall back on the same rhetorical and structural techniques, so the "new style" is strictly surface-level. In fact, it's getting worse over time -- especially in Claude from Opus 4.6 --> Opus 5.

I suspect that the same thing is happening here -- as other comments note: lack of distinctiveness, uniformity -- but that the low information content of posters prevents it from becoming immediately obvious.

A_D_E_P_T··on Stripe Withholding Balance of $100000
If payout is stuck for "2-3 business days," you still have account access, and you're still capable of accepting funds, you should probably just relax. Odds are that everything will be fine by Tues or Weds. If the transactions are legitimate and you can prove it, then the odds that they'll confiscate your money or hold onto it for an extended period of time are very, very low.

Not a bad idea to set up a secondary payment processor for growth, though!

A_D_E_P_T··on Anthropic builds wet lab for AI-driven drug discovery
This has got to be some kind of "how could things possibly go wrong?" AI Safety publicity stunt.
A_D_E_P_T··on Will Taiwan Just Give Up?
There's movement in that direction. "Beijing hails Taiwanese scholar’s rare proposal on reunification," from two days ago: https://archive.ph/2VSMT

It's hard to view it as a bad outcome next to the alternative, though.

Also the birthrate in Taiwan has absolutely cratered. TFR is 0.69.

> https://www.straitstimes.com/asia/east-asia/price-of-parenth...

I believe that this is among the lowest vales ever recorded, anywhere. It's simply an objective fact that Taiwan is not a vibrant or healthy society. The status quo obviously isn't working for them.

A_D_E_P_T··on Watch AI materials-science and bioscience abilities closely
Hah, well, you must have never encountered LessWrong before. It has been like that for well over a decade, in fact it's probably approaching its second decade...

One thing you've got to give them credit for is that they've been talking about AI -- about the present moment, really -- from the beginning. Nick Bostrom, who wrote "Superintelligence" about ten years ago, has always been in their orbit.

A_D_E_P_T··on Israel accused of eliminating living conditions for Palestinians in West Bank
At this point it couldn't be more obvious that there's something terribly rotten in Israeli society.

The "respectable" political center is afraid to talk about it, though. This only pushes normal people with well-calibrated senses of justice to the extremes of left and right.

Much more action needs to be taken against Israel, and, for straightforward moral and ethical reasons, this really should be a mainstream political position. If they act as though they're a pariah state who can go it alone, they ought to be a pariah state forced to go it alone.

A_D_E_P_T··on Predictive intelligence to anticipate anything.
I jailbroke it a little bit. It's running Gemini 3.7 Flash.

Interestingly, I asked it a question, and asked 3.8 Flash the same question (as I no longer have easy access to 3.7 Flash) and the answers were NOT identical, as 3.8 Flash just mindlessly repeated whatever odds were live on Polymarket.

So then I asked 3.8 Flash to disregard the prediction markets and run its own analysis, and it produced essentially the same answer as the Prior.Chat interface.

Long story short, as far as I can tell it's running 3.7 Flash with a decent custom prompt. You'd probably get materially better results if you just asked a stronger model like Fable or Astra for a prediction...

A_D_E_P_T··on 'Fingerprints' inside the Sun could reveal if it once swallowed a planet
> Their results favour a scenario in which the young Sun engulfed a super-Earth around 5–10 times the mass of Earth

Given the state of the early solar system, it would be surprising if the Sun didn't swallow a lot of metallic and carbonaceous rocks and pebbles. But it seems unclear to me how the researchers would distinguish between one super-Earth and >10^10 smaller rocks? Chemically I feel the two would be largely the same.

A_D_E_P_T··on The worst spam emails: iLands AI agent hustle
It's the presumably human x.com screenshot where the paragraphs are repeated. The LLMs, though they all seem to write in that characteristic clipped tone which one can now spot from a mile away, really wouldn't make such a mistake.
A_D_E_P_T··on Diesel prices in U.S. top $6 a gallon for first time
That needs an update: Winning wars, if you win in a clean and non-Pyrrhic way, is usually good for the economy.

Losing at war is universally bad. And long stalemated forever wars can be just as bad for both sides.

A_D_E_P_T··on [dead]
You write: "Nazi, deportation and remigration party wins in Germany"

The actual headline is:

> "Far-right AfD wins vote in Germany’s Saxony-Anhalt state: Exit polls"

Portraying every right wing party as "Nazi" is a grave mistake. The AfD, for all their many faults, are not Nazis. Simply consider the fact that they are led by a lesbian woman married to a Sri Lankan immigrant. In the 1930s, this would have been denounced as peak Weimar degeneracy, and the pair of them would have been the very first to the camps.

The actual danger is that by calling tepid right-wingers Nazis, you get them saying, "Well, if they're going to call us Nazis anyway, maybe we ought to go harder. We have nothing to lose if we take a more radical stance."

Semantic drift ---> More political polarization. It's a real problem.

A_D_E_P_T··on Tell HN: Anthropic just reset usage, including Fable
Yeah, it's live right now in my Pro subscription, but I need to toggle either "Pro Mode" or "Work Mode" to activate it.
A_D_E_P_T··on Tell HN: Anthropic just reset usage, including Fable
Yeah, I just noticed that myself. I figured it's because GPT-6-Astra was just released, so it's a way of saying, "please don't cancel right now -- or at least take these tokens and run some comparisons for yourself."
A_D_E_P_T··on Ask HN: Why were OpenAI, Claude, and Grok simultaneously down?
https://archive.ph/3zcMG

It was extremely weird... and if it was a load thing they probably would have explained it by now?

A_D_E_P_T··on Dextroproporphan: An Analogue for a Better Dextromethorphan?
If you're OP and you're considering clinical trials... well.... I hate to be the bearer of bad news, but you've got another problem. The compound you've described was also described in the scientific literature (back in the 90s) and it's in at least one patent application:

> https://patentimages.storage.googleapis.com/37/75/89/204db4e...

See claim 16 and "compound 102." So, unfortunately, fatal prior art exists. (On the bright side, those scientists arrived at exactly your hypothesis.)

My own hypothesis, that deuteration would be useful, is also in the patent literature: https://patents.google.com/patent/US10406155

As is the deuterated isopropoxy!

I'd suggest an alternative, but merely mentioning it here might constitute prior art. If you're interested in discussing, my email is in my profile.

A_D_E_P_T··on Dextroproporphan: An Analogue for a Better Dextromethorphan?
The CYP[XYX] enzymes process just about everything that isn't a starch, a protein, or a lipid. So whenever you eat anything, something in that food is going to interact with a CYP enzyme somehow -- even the plainest white bread contains acrylamide, which is metabolized by CYP2E1 and others.

Point is, there's a huge laundry list of drugs and dietary chemicals that inhibit or induce CYP enzymes. That Wikipedia list is far from complete.

These guys claim that something in goldenseal can inhibit CYP2D6 by >50%: https://pmc.ncbi.nlm.nih.gov/articles/PMC2562884/

...If there's a phamacological interest in CYP2D6 inhibitors, whatever goldenseal constituent is responsible (and it's assuredly a small molecule that is metabolized by CYP2D6, because that's how these things work,) can be isolated and improved upon. This is relatively simple and predictable in comparison with making new drug analogs and estimating their effect.

So if I were OP, I'd focus more on CYP2D6 inhibition as the surer thing.

A_D_E_P_T··on ChatGPT Is Throwing 404
On my end Kimi K3 is still up, and Gemini is up at gemini.google.com. Claude and ChatGPT are both down.
A_D_E_P_T··on Fish bad, sugar good, and other medieval ideas about food
> "Fish Bad, Sugar Good"

> "Similarly, some fruits were safer than others: citrus fruits improved digestion, especially in hot weather, while raisins were good for the stomach, lungs and liver. And nuts were nutritious and good for the brain."

My impression is that refined sugar was extremely rare in medieval times and into the renaissance. But everybody loved raisins, and they were used extensively, even though they were usually imported from Italy and other Mediterranean regions. Various forms of raisin bread (often with dried figs, citrus peels, and currants,) were the definitive medieval festival/feast-day food. Their distant descendants are the stollen, panettone, etc., still popular as holiday treats in Europe.

Certain regions of medieval Germany (and presumably France) grew grapes but had to resort to ingenious methods to dry them into raisins: https://www.culina-vetus.de/2026/08/27/raisins-in-a-cold-cli...

A_D_E_P_T··on Can I opt out of my input or output data being used for training?
EU companies can download GLM or Kimi-K3 and get way better performance, though? I don't see any case for using a Mistral model when much better open models exist.
A_D_E_P_T··on Dextroproporphan: An Analogue for a Better Dextromethorphan?
Isn't CYP2D6 inhibition the surer and more convenient route? Adding a bulky isopropyl group might reduce affinity at sigma1 receptors. Think about it: Going from 3-methoxy to 3-hydroxy makes for a very different drug. Wouldn't you expect that going from 3-methoxy to 3-isopropoxy would also make for a very different drug -- and not just "the same thing but more resistant to first-pass oral metabolism"?

Selective deuteration seems like another potential option with good prospects. Here it reduced CYP2D6 metabolism in a broadly similar case: https://pmc.ncbi.nlm.nih.gov/articles/PMC8724172/

Deuteration in a position like that is so simple that you can do it at home.

A_D_E_P_T··on LLMs and Self-Referentiality
> The big, old ideas about intelligence that ended up basically vindicated were the ideas about how intelligence is about prediction, and prediction is about compression, and compression is about finding better and better upper bounds on Kolmogorov complexity. Not the self-reference stuff.

> What’s left? Consciousness and subjective experience of course remain extremely mysterious. For all we know, Hofstadter could be right that those have something to do with self-reference. (For all we know, even Penrose could be right that they have something to do with exotic physics accessible to biological brains but not digital computers!)

I'm not convinced that this is the case. Biological entities have a spatial-temporal sense, a memory, and, alongside those, a predictive/simulating function in space and time. A cat who knows where it is in the world at a given time, sees a mouse, and predicts where that mouse is going to go, is essentially performing a physical calculation, much as you do when you attempt to catch a ball. The more [powerful + general] the predictive function, the more intelligent a given biological entity is. Past a certain threshold, the predictive function becomes capable of broad abstraction. This enables man to acquire, process, store, represent, and continually re-acquire (in a Bayesian sense) knowledge that is external to that man's subjective existence. This is pure calculation in the strictest sense, and doesn't require exotic physics or self-reference beyond a sense of being bounded in space and time, which is to say a reference point.

All that aside, Hofstadter and Chomsky have been pretty hard-hit by LLMs!

A_D_E_P_T··on Konrad Zuse Museum shutting down due to lack of funding
It has been said -- I believe by Stephen Wolfram, who owes a very great deal to Zuse -- that the universe is "computationally irreducible." In other words, and particularly in this case, that every particle would need to be simulated by a similar particle or equivalent mathematical representation. And thus that any (quantum) computer running a simulation of our universe would need to be the size of the universe, and would require the energy of the entire universe.

In any case, the laws of physics are not yet settled, and determinism is still a very live option. The notion of a continuum is, to me, even more bizarre than the notion of a deterministic universe -- and is also more bizarre than Everett's notion of a "wave function of the universe."

A_D_E_P_T··on Konrad Zuse Museum shutting down due to lack of funding
Zuse is probably the most under-appreciated thinker of the 20th century. Everybody's heard of Claude Shannon, von Neumann, Feynman, and all the rest of 'em, but Zuse's enormous contributions have received rather little attention.

To summarize a couple of the highlights:

In 1936 Germany, Konrad Zuse began developing the Z1, a mechanical calculating machine programmed via perforated 35mm film. By 1937, he had broadened its architecture toward universal digital computation in the vein of Alan Turing. Although completed in 1938, the machine struggled with mechanical inaccuracies -- the same challenge which plagued Charles Babbage's much earlier work on the Difference Engine. Zuse overcame these mechanical hurdles in 1941 with the Z3, an operational digital computer capable of universal computation.

In fact, the Z3 was the world's first working machine physically capable of universal computation. The first real computer.

Zuse was also one of the great philosophers of the 20th century. In a book he published titled "Calculating Space," He was the first to surmise that our universe is computable by a deterministic computer program. Every "simulation hypothesis" and "mathematical universe" theory is downstream of Zuse. (One could argue that the Pythagoreans and neo-Pythagoreans, who held that the world was at root mathematical, were truly first, but Zuse made important and original contributions nevertheless.) It's important to note that, to this day, there are no physical observations which contradict Zuse's notion, as he wrote it, of a computable universe.

I would think that the German state or EU should provide funding for his museum. If that's not the case, it's a shame.

← PreviousPage 2 of 34Next →