Google was not marginally better Yahoo, their implementation of Markov chains in the PageRank algorithm was significantly better than Yahoo or any other contemporary search engine.
It's not obvious if a similar breakthrough could occur in AI
I'm fairly certain that literally every doctor tells literally every patient this. Mine sure did. It's also on the handout your pharmacy gives you, the website for the drug, and in fine print at the bottom of most commercials.
Nobody expects a single does of ibuprofen to cure your headaches for life. Similarly, this doesn't "fix" your biochemistry for life, you have to take it in order for it to work.
I've been on Zepbound for ~25 weeks now and one of the first things my doctor told me was that this was a lifetime drug. He pointed out that I had a choice, of lifetime drugs. I needed to do something to get my cholesterol under control. That meant start statins now and likely add blood pressure meds and diabetes meds in the next 2-5 years. Alternatively I could start Zepbound, which would likely address all 3 and result in better quality of life in the next year. So far it seems like it is doing exactly that.
I'm sure some people are going into this without that knowledge, but people are being told this is a lifetime commitment. What you don't see a lot of is why people stop taking it. There's some cases of people losing and then stopping, but the majority are because insurance is forcing people off of it. Just look at the recent CVS Caremark forced switch from Zepbound (2nd Gen) to Wegovy (1st Gen) in July.
> Just look at the recent CVS Caremark forced switch from Zepbound (2nd Gen) to Wegovy (1st Gen) in July.
This is something I wasn't aware of, are you on 2nd gen Zepbound then?
My wife has been plateaued at an undesirable weight and has been wanting to try this, however, the VA refuses to support it regardless of the fact she fits their guidelines and requirements to receive it. They recently banned it due to costs.
I prefer her to use Zepbound if we can get it, the question is how? We refuse to use the alternative methods where the drug is hand made to be equivalent to them as that seems very sketchy.
> I suspect that this info is intentionally down-played so that it doesn't affect sales.
Yea, hard to figure out why drug companies keep producing "Faustian bargains" in our current system. What galls me is people assume the best for new drugs instead of forcing the _for profit_ entity to prove it's actually safe and useful.
To the extent that, reliably, the first comment on these posts on Hacker News are some wishy washy anecdotal emotional blackmail garbage that completely obfuscates the point and runs direct interference for these large profitable organizations.
To the extent that it's hard to believe that these posts even on this tiny corner of the internet aren't bought and paid for. We live in a society that cherishes organized crime and denigrates hard work. I would not look forward to "new drugs" in this regime.
Novo Nordisk Foundation is a nonprofit but Novi Nordisk the pharmaceutical company is very much a for profit entity - it’s publicly traded on several stock exchanges including the NYSE. The foundation owns most of the voting shares but only about a quarter of all outstanding shares.
> people aren't aware of how wrong they can be, and the errors take effort and knowledge to notice.
I have friends who are highly educated professionals (PhDs, MDs) who just assume that AI\LLMs make no mistakes.
They were shocked that it's possible for hallucinations to occur. I wonder if there's a halo effect where the perfect grammar, structure, and confidence of LLM output causes some users to assume expertise?
Computers are always touted as deterministic machines. You can't argue with a compiler, or Excel's formula editor.
AI, in all its glory, is seen as an extension of that. A deterministic thing which is meticulously crafted to provide an undisputed truth, and it can't make mistakes because computers are deterministic machines.
The idea of LLMs being networks with weights plus some randomness is both a vague and too complicated abstraction for most people. Also, companies tend to say this part very quietly, so when people read the fine print, they get shocked.
> I wonder if there's a halo effect where the perfect grammar, structure, and confidence of LLM output causes some users to assume expertise?
I think it's just that LLMs are modeling generative probability distributions of sequences of tokens so well that what they actually are nearly infallible at is producing convincing results. Often times the correct result is the most convincing, but other times what seems most convincing to an LLM just happens to also be most convincing to a human regardless of correctness.
> In computer science, the ELIZA effect is a tendency to project human traits — such as experience, semantic comprehension or empathy — onto rudimentary computer programs having a textual interface. ELIZA was a symbolic AI chatbot developed in 1966 by Joseph Weizenbaum and imitating a psychotherapist. Many early users were convinced of ELIZA's intelligence and understanding, despite its basic text-processing approach and the explanations of its limitations.
Its complete bullshit. There is no way anyone ever thought anything was going on in ELIZA. There were people amazed that "someone could program that" but they had no illusions about what it was, it was obvious after 3 responses.
Don't be so sure. It was 1966, and even at a university, few people had any idea what a computer was capable of. Fast forward to 2025...and actually, few people have any idea what a computer is capable of.
My experience, speaking over a scale of decades, is that most people, even very smart and well-educated ones, don't know a damn thing about how computers work and aren't interested in learning. What we're seeing now is just one unfortunate consequence of that.
(To be fair, in many cases, I'm not terribly interested in learning the details of their field.)
If I wasn't familiar with the latest in computer tech, I would also assume LLMs never make mistakes, after hearing such excited praise for them over the last 3 years.
It is only in the last century or so, that statistical methods were invented and applied. It is possible for many people to be very competent at what they are doing and at the same time be totally ignorant of statistics.
There are lies, statistics and goddamn hallucinations.
Have they never used it? Majority of the responses that I can verify are wrong. Sometimes outright nonse, sometimes believable. Be it general knowledge or something where deeper expertise is required.
I worry that the way the models "Speak" to users, will cause users to drop their 'filters' about what to trust and not trust.
We are barely talking modern media literacy, and now we have machines that talk like 'trusted' face to face humans, and can be "tuned" to suggest specific products or use any specific tone the owner/operator of the system wants.
Magnacut is given the same rating as D2, 3V, VG10, and "440" (presumabbly 440C). Elmax, M2, and Cruwear are rated higher.
CATRA testing by Magnacut's creator[0] suggests D2, 3V, VG10, and 440C should have a lower rating than Magnacut for edge retention, while Elmax, M2, and Cruwear should be the same. 5 probably is a reasonable rating for Magnacut though when Maxamet is 10.
Larrin's ratings handle this by not having a fixed scale; REX 121 gets a 12 for edge retention. The rating seems to be roughly CATRA cards cut divided by 100.
Did you know you can buy an Oxford Nanopore sequencer for about $2000 euro? It sits on your desk and it will sequence DNA or RNA at about 450 bases per second
It somehow grabs a free strand of DNA\RNA and pulls it through a specially constructed pore on a membrane. It reads the voltage change as it goes through because each type of base (ATGC) changes the voltage differently, and interprets the sequence and outputs the data in a file.
Doesn't even need PCR amplification because it reads a single strand.
I thought it must be fake because it sounds like something out of science fiction but Nature did an article:
What's the accuracy of this compared to commercial sequencing services, and what's the actual TCO? Can 20 people do a group buy for one machine, and use it to get truly private DNA sequencing for $150 ($100 capex + $50 for consumables)?
> You actually don't have to power it down. If you hold the power and volume buttons for 2 seconds and reach the "slide to power off" screen, the phone is already hard locked. You then always have to enter the passcode to unlock it.
Iphones have 2 states when it comes to encryption:
Before First Unlock (BFU) - everything is encrypted. The most difficult state to hack.
After First Unlock (AFU) - data isn’t fully encrypted. Maybe it's for performance reasons. In this state exploits exist which police can use to get data.
Your suggestion of getting to the 'slide to power off' screen does NOT hardlock the phone (it does not put it in BFU).
It just means it requires a passcode. However, since it is in AFU mode, data can be exfiltrated with the right tools.
So if you are _not_ someone with an antisocial personality disorder and you go to therapy like couples therapy you basically learn that you need to be kind and thoughtful even if your partner is not and compliment them and it basically rubs off on them. It's also the way you make friends etc...
It's the basic psychological principle that you need to give to the world what you wish to receive. It is counterintuitive.
I'd guess that this would be the treatment program for APDs as well except that the APD may short circuit the reciprocal reward program.
Eh? (P(X|Y)>P(X) iff P(Y|X)>P(Y)) whenever all the relevant probabilities (P(X), P(Y), P(X and Y), etc.) are strictly positive and strictly less than 1.
Now, X might be much stronger evidence for Y than Y is for X, but if Y is evidence for X, then (assuming the previously mentioned assumption) then X is evidence for Y, even if only very weak.
I think in the coming years we’ll agree that the notion of a psychopath as described/defined by pop science was completely off base and made up mostly to discredit opponents in one way or another. First, however, we have to find a valid substitution so we can swing to the next branch before the plebs catch on that we were making shit up.
Some of it is probably real, like repeat offenders of predatory crime. If you've done multiple violent robberies there's probably something in your brain off that allows you to do all that.
There's a predatory ecological niche in every society. Very common in nature too. Altruism is less common than predation in nature, it seems.
Though in the workspace, you may be onto something.
Why wouldn't there be psychopaths in the workplace?
The way I see it there's a selection bias at play here. The kind of psychopaths who are repeat offenders are the ones who get caught, so of course we notice them.
We don't notice the ones who don't get caught.
That doesn't mean that they're any less ruthless and destructive to society -- they're just lucky / more functional.
In my mind I view white collar crime to be the greatest evil in society because it is the thing that enables all of the petty/violent crime that we abhor. Part of the reason that it's so evil to me is for that very reason, and that the perpetrators of it are able to pass themselves off as functional individuals who participate in society like you and I do, and are able to use the system against itself to perpetuate their crimes.
The mark of the actual master is that almost no one can identify them, and the people who can are so rare as to probabilistically not appear in smaller companies. They “run” the company without the appearance of doing so.
If someone does appear who can identify them, they are quickly exited from the company via various means depending on the individual: pair them with a particularly distasteful partner/manager, make their job impossible via passive means, or enable them to humiliate themselves.
The only real symptom is that there is no one around them that competes with them in their self-perceived core competency, and they never seem to put a foot wrong. Once the are in a high position, anyone who complains about their behavior is said to be attempting to climb.
If you've ever met a psychopath and seen their true colors, you will undoubtedly be convinced it's a real phenomenon. I have had the unfortunate experience of dealing with one myself. He showed all the hallmark signs. He urinated the bed frequently. He harmed animals when he was young. Specifically, he swung his pet gerbil around by the tail until the tail was severed. He threw a baby rabbit against a stop sign and killed it while his poor younger brother watched. He has grown up to be incredibly sadistic to the point where everyone I know has completely cut him off. I hope you never run into one. When you are talking to someone you believe to be decent, and you eventually realize what they are, what you're dealing with, it is a truly chilling experience.
It's true, a direct link hasn't been proven, and I was debating even mentioning that. Obviously that specific tendency alone isn't indicative of a homicidal individual. But I would speculate that it has something to do with dysfunction/dysregulation of the autonomic nervous system.
Yes, but that is the absence of evidence, commonly called on when the science hasn't been done.
If a theory/hypothesis remains not proven once the science has been repeatedly done in numerous attempts to prove it, very often it does mean it is wrong.
It still seems like it could be correlated, but it shouldn't be used as a diagnostic measure or taken as a sign of something. It's more likely to be a sign of abusive parenting, but abusive parenting is also correlated with the psychopath behavior so it all ends up being correlated with each other.
The replication crisis was recently discovered; this does not mean older science does not have the same challenges. Generally the problems that led to the replication crisis are worse the further back you go in time
3 letter hotkeys seem to work well - long enough to be unique without overlapping real words.
for instance, when I type ddd it automatically stamps this:
20250811 10:57 AM
then I type my note and can look back at what time\date it was.
sometimes I just need the date so that is dds (date-date-short) which gives
20250811
occasionally I just want the time so that's ttt
11:02:02 AM
I have many other 3 letter codes using Autohotkey to bring in frequently typed things too, useful in emails and such.
Since I think through typing, autohotkey has been a QoL helper.