HNHacker News
TopNewBestAskShowJobs

hirvi74

1,987 karma · joined August 11, 2020

submissionscomments
hirvi74··on “Next-token predictor” is the wrong mental model for LLMs
Sure, I get the gist of the article. I have never liked the reductionist argument that LLMs are nothing more than next-token predictors. By that rational, the human brain is really not that much different. When I am having a conversation with another person, I do not usually have every word I will respond with stored in my limited working memory. My output is often predicted based on the previous word I spoke.
hirvi74··on Google Antigravity TOS: 3rd party usage can get Google account suspended
Couldn't one just make a separate Google account for the models/AI products? I usually have multiple accounts with most providers, multiple banks, multiple brokerages, etc.. It really is not that difficult to manage.

Though, I only have one HN account, fwiw.

hirvi74··on Gloria Steinem has died
Why do people always have to pinpoint some flaw, issue, or mistake in order to tarnish people that have achieved something great?

Steinem was a human, not a paragon of perfection. She least used some of her time and energy on this Earth to try to make the world a better place.

hirvi74··on The asteroid currently hitting front end web development
That is why I am very thankful for my job. We are not really allowed to use LLMs for much of anything.
hirvi74··on Mushroom hunting with LLMs: what can go wrong?
I had a friend that was quite passionate about foraging for mushrooms while we were in high school. He eventually went on to receive a PhD. in mycology. My old friend as well as everyone else I knew that liked to forage mushrooms were... uh... not looking for "flavorful" mushrooms to say the least.
hirvi74··on Hang on to Your Firefox
I'm good with my Safari, but thanks.

(Though, I do use FF on Windows).

hirvi74··on Tim Cook Sold Out Steve Jobs (2025)
If anything, those aspects were probably highly correlated to his success. Our economy runs on a system that does not prioritize nor reward empathy.
hirvi74··on Reverse engineering my ADHD test
I think there is a conflation between attention and attention to details. I understand attention control can vary depending on the engagement value of a task. I have experienced the condition my entire life, I am well aware.

I am not aware of any data that supports being interested in a task lessens the likelihood of making mistakes. If you have any data, I will gladly read it.

hirvi74··on Claude Fable 5.1 and Claude Mythos 5.1
I am surprised Anthropic can use their models accurately solve this issue?

Are different services for different users based on geolocation really that difficult? I thought a lot of services operated like this already.

hirvi74··on Reverse engineering my ADHD test
> People with ADHD are differently wired in their brain

I am not sure that is objectively true in the sense that people think. If so, the condition would have a much more objective diagnostic process. I have been professionally described as ADHD, had an fMRI (unrelated reason), and my results came back perfectly normal.

I know n=1, but the data doesn't necessarily suggest that ADHD is some kind of organic, neurological disease.

hirvi74··on Reverse engineering my ADHD test
> ADHD often comes with a high attention to detail

Per the DSM-5, one of the core symptoms of ADHD:

"often fails to give close attention to details or makes careless mistakes in schoolwork, work, or other activities

hirvi74··on Reverse engineering my ADHD test
There might be a correlation, but I doubt causation. I would argue people with ADHD, like myself, tend to end up in situations where extreme pressure and chaos is more common. Thus, we have more exposure than the average person. I imagine it can be a trainable trait in many people.
hirvi74··on Google Has Removed MV2 Extensions from the Chrome Web Store, Including UBO
I would enjoy automatically generated subtitles more if they weren't wrong about 10% of the time.

I constantly get something like "Today, we are going to discuss which model is better at programming. Is it an tropic's latest model, Cloud Fable? Or is it, Open AI's Chat G pity?"

It's as though Google hired my dyslexic ass to write the subtitles. It trips me up even more because I can't tell if I misread something or the chat messed up sometimes.

I cannot comment on translation or dubbing. English is the only language I can somewhat understand.

hirvi74··on Google Has Removed MV2 Extensions from the Chrome Web Store, Including UBO
Okay, Google bought YouTube earlier than I remembered (2006). For some reason, I remembered the purchase being somewhere around 2009-2012.

So, if I may change my opinion, early YouTube was better, despite being owned by Google. I feel like the platform has been atrocious in the last decade or more. Constant clickbait, a new NordVPN or Raid Shadow Legends ad every 15 minutes, politics being injected into every other video, creators trying to game the algorithm to escape the rat race, the obnoxious algorithm itself, and now I am getting an absurd amount of LLM generated videos.

I would not dare claim there is not great content though. I am listening to a YouTube video as I type this. Though, I have been listening this particular creator since 2006, if that tells you anything.

Nevertheless, I cannot help but feel YouTube is just getting the slow squeeze of death -- like everything Google wraps its tentacles around.

hirvi74··on Evidence of Fraud in an Influential Study About Procrastination
Does some sort of batched style of research already exist? By batched, I mean a study where n number of teams all agree on a single methodology, and then conduct the same experiment simultaneously. Each team could then publish their results independently and the teams could publish one aggregate.

I know meta studies attempt to do this in a way, but they can sometimes have their own problems too.

I cannot recall coming across this style of research in much of what I have read as a hobby, but maybe it is more common in some domains than others? I am not smart, and there is nothing new under the Sun, so surely something like this exists, no?

hirvi74··on Google Has Removed MV2 Extensions from the Chrome Web Store, Including UBO
I'm with you. I've been riding with Safari for almost 20 years now. It's been nothing but great for me. I cannot think of a time I have used a website that did not function properly in Safari.

Prior to the reintroduction of UBO (lite), I feel like Safari helped me reduce my trash content consumption. Page with tons of ads? Not worth visiting. Youtube keeps playing ads? My frustration is a sign I need to get off YouTube and do something else. I do not need nor want to mange tens of plugins. I like to keep things simple and slim. So, perhaps that is one dealbreaker for some?

Even if I believe Apple hung the Moon, I still seriously recommend Safari to anyone using Apple devices.

hirvi74··on Google Has Removed MV2 Extensions from the Chrome Web Store, Including UBO
It was a better platform when it was its own company back in the day.
hirvi74··on Casey Muratori – The Root of the Root of All Evil – BSC 2026 [video]
I have no idea what the parent meant by, "My guess is that your ASM is inflected by structured programming everywhere."

(I am the GP)

Your response was far better than mine. If those instructions were executed more than a couple of times and the constraints/assertions were identical, I'd absolutely throw them in a function. I'll handle what I need to before/after the function.

I would never branch to some coincidental label. That is why I love ASM so much. If you are sloppy or lazy, you will most likely be punished severely for those choices.

(Tangential, but when I first learned ASM in college, I felt like I learned more in that one semester than all the sum of all classes in my entire degree.)

hirvi74··on Casey Muratori – The Root of the Root of All Evil – BSC 2026 [video]
To be honest, I do not have much experience with C nor BASIC. I only know the extreme basics of C, and I have never written a line of BASIC in my entire life. My first real exposure to programming was in high school (Java). From there, I basically went from high-level languages -> x86 -> high-level languages again -> starting to play with AArch64 in an attempt to but some grooves back in my smooth brain.

> My guess is that your ASM is inflected by structured programming everywhere.

I am not certain what you mean.

> if you go back 60+ years there is some scary code that even today's optimizers probably wouldn't emit because it's too crazy.

Why was this as common as the allegations lead me to believe? Was it just a product of time? No better way to handle such logic at the time?

As for your example, I know it is intended to be interpreted trivially, but I just want to clarify I understand you correctly. Is the issue due to locate-horizon depending on logic from a completely orthogonal operation?

If so, I see the red flag immediately. However, I do not see an issue with creating a separate function. I'd just have to be damn sure H was not storing anything used for some other upcoming operation prior to the function call, I'd check F and C for valid state (if necessary), take other potential side-effects into account, etc.

It feels a lot like the DRY advice being abused in higher-level languages, you know? I agree with expression, "duplication is better than the wrong abstraction."

hirvi74··on Doctors are finally learning to manage antidepressant withdrawal
Addictions require harm in one or more areas of your life, and the compulsive use or inability to stop despite the harm(s).

I imagine many diabetics crave insulin when they cease its usage.

hirvi74··on Doctors are finally learning to manage antidepressant withdrawal
My understanding is that addictions have to have an element of psychosocial or physiological harm. Particularly compulsive usage and the inability to despite the harms is when something becomes an addiction.

Quitting coffee might give someone headaches, bit more tiredness, difficulty concentrating, etc.. But that is just the body trying to achieve homeostasis.

Many diabetics would have horrible withdrawal effects from ceasing their insulin usage -- a substance they cannot live without. They take it everyday and their life depends on it. But many live perfectly normal, functional lives. They do not break into cars to steal radios to pawn off for more insulin.

However, jeopardizing your family’s future by taking out a second mortgage on your home just so you can keep gambling -- while you’re already drowning in debt -- and telling yourself, "If my wife finds out, she’ll divorce me and take the kids. But this is the time I win big. This time will be different."

That, right there, might be the purest form of addiction there is. All without any physical dependence at all. Gambling addiction also has the highest suicide rate amongst all addictions.

hirvi74··on Casey Muratori – The Root of the Root of All Evil – BSC 2026 [video]
So, I have never seen anyone actually use GOTOs, so maybe a lot of the stigma comes from excessive, poor usage or something.

However, having written a good chunk of ASM in my life. I don't think jumps or branches are really that hard to follow. Jumps/Branches and GOTOs specify the next location. It is not as though one has to guess where.

It's not the arrow, it's the archer that is the problem.

hirvi74··on Casey Muratori – The Root of the Root of All Evil – BSC 2026 [video]
> not what is demanded nor paid or appreciated.

I suppose it is a matter of what one works on, but that has been consistent in my career.

I have seen many be rewarded for choosing what I would deem intentional, gross negligence. But hey, those devs get work finished faster, and that's all that matters to the non-technical folks.

hirvi74··on No AI Fridays
> ...cause you to accumulate cognitive debt, make you less engaged with your work, negatively impact your critical thinking abilities, and hamper your skill formation.

My employer gifted me all that years before LLMs were released.

hirvi74··on Doctors are finally learning to manage antidepressant withdrawal
I say this as someone with ADHD:

The issue is that ADHD is not real in the organic sense. Well, at least the evidence is scant. It's more of a description than a disorder. It's not something one "has."

I like to compare ADHD to racial categorization of people. Race is not real in a biological sense, but it's definitely real in a social sense.

I am probably butchering what I am trying to say so, just read this if you want to understand someone(s) that can communicate eloquently.

https://pmc.ncbi.nlm.nih.gov/articles/PMC9189308/

hirvi74··on Doctors are finally learning to manage antidepressant withdrawal
That is not a characteristic of ADHD. Th most damaging thing to ADHD is the myth that the medications affect people with ADHD differently. It is the most pseudoscientific piece of information that is commonly repeated.

As for tolerance, it absolutely happens. Most people do not stay on the medications for a long enough time to really feel the tolerance to begin with. Sure, in a clinical trial of four to six weeks, it might not be noticed. I've been on them for well over a decade, and it's a night and day difference.

hirvi74··on Doctors are finally learning to manage antidepressant withdrawal
But how does one know? According to some studies, spontaneous remission happens in about half the cases of depression in a given year.

It's happened to me plenty of times. That's why I refuse to medications. All I have to do is just wait it out.

I am not claiming no one gets any benefits, rather how can one be certain the benefits are actually not from something else? If the goal is to improve, then I suppose it does not truly matter. But I just like to understand things when I can.

hirvi74··on Aphantasia Beginner's Guide
I meant definitely proven with in a realistic sense. Smoke causes lung cancer, albeit only something like 20% of smokers get lung cancer. It's been proven, but not in the mathematical sense of the word.
hirvi74··on Aphantasia Beginner's Guide
One, a study with more than 32 people. Two, replications with equally strong observations.

I've read a lot of research on fMRIs and mental disorders. Something like this pops up every other day. Not a single one has really study the test of time ion terms of diagnostic utility.

hirvi74··on Judge rules Trump administration’s blacklisting of Anthropic was illegal
You're not wrong. The United States has a legal system and not a justice system. Having spent many years of my life involved 'adjacently' with the system, I'd say it's actually a rigged, sadistic system.
← PreviousPage 4 of 34Next →