HNHacker News
TopNewBestAskShowJobs

hanselot

238 karma · joined January 11, 2017

Its important to understand that i don't see myself as a reflection of my past. I leave it all there because I'm not afraid of learning from my history. So if you engage with me, take that into consideration. Don't train yourself against an old dataset.

Love and respect. Xænu

submissionscomments
hanselot··on Linux has achieved a 3% desktop market share
I was recently surprised to find installing battle.net with Steam Proton is not only possible, but dead simple, and actually worked. I had been trying for days using the lutris installer, but just shit keeps breaking endlessly. At some point I compiled lutris into a venv just to try and solve all the dependency shit that was lurking in there.

But lutris as it stands now breaks everything almost every time I use it. I had steam unlaunchable for over a week because of something lutris was doing implicitly. Now running steam directly off the CLI seems to be working for 90% of the games I play (which isn't a very long list).

Only thing I doubt I will be able to figure out is stuff like fortnite or anything that uses easy-anti-cheat (since their devs give a fuck about linux and effectively believe we are all hackers trying to ruin their game).

That's all I got, don't bother with Lutris unless you have an insane amount of time to debug which component is missing.

hanselot··on Let the internet be grimy
But once you reach 109 times it rolls back to being coincidental right?

Those qui know, know.

In any case, lets not be deemed offensive for harboring thoughts of rebellion against the overlords lest we be yeested from the platform for attempting to pique the curiosity of those not entirely lobotomized by the oppression machine yet. Such fallacious behaviours shall be left without scrutiny while we all dance together under a rainbow of landmines, minimum wage and proxy wars funded by alphabet companies ever-seeking to exercise greater and greater control of the human mind at scale.

Yes, let's strive not to offend others, for that way we will surely make the most progress, ignoring all prior knowledge and our very evolutionary mechanisms so we don't make lines on a page that induce rage in the untrained mind.

Those that can't hold a thought without judgment, can't think critically, yet they are the ones we decide to put in charge of our social narrative?

I wonder WHO decided that was a good idea.

hanselot··on Rejected GitHub profile achievements
Congrats on earning the "Adding Value" achievement:

Pointing out something obvious, then having someone else point out that what you pointed out is obvious, then defending the obvious thing that you pointed out by stating that it in fact is not obvious

hanselot··on How many people have ever lived on Earth?
I don't like to comment much anymore, but I feel like I have spent a substantial portion of my life on this topic, and consequently can direct people to some resources that can bridge the knowledge gap a bit in this regard.

>delay it by decades

Yes, but we aren't solving the problems, we are dealing with symptoms. The entire industry is currently treating old age diseases as diseases, and not as symptoms of bodily decay. This is covered extensively by the work of Dr Aubrey de Grey (yes, that one), in any case, he documents 7 different ways in which the body accrues damage over time, and is actively researching in all these fields to find methods to mitigate, and reverse said damage.

To aid in this regard, feel free to visit the SENS research foundation website at https://www.sens.org/

This approach is very different from the one currently deployed in the field afaik. The idea is that the human body, like a car, can be kept in pristine condition if these 7 sources of deterioration is reversed/addressed.

It is really quite fascinating. I would advise you to go check out the healthspan discord on https://discord.gg/sCwyPXu

There are a number of highly motivated individuals actively researching and following the SOTA as well as excellent news aggregation.

This is the initial reason I got interested in this in the first place: https://www.youtube.com/watch?v=AvWtSUdOWVI

One of Dr de Grey's quotes that I find quite compelling is:

"There is no difference between saving lives and extending lives, because in both cases we are giving people the chance of more life."

hanselot··on ChatGPT: Fear Litany
Sure.

https://chat.openai.com/share/d9284035-0820-4a5c-9e54-a7c411...

https://chat.openai.com/share/bce6ee2f-604a-413a-b14a-7145b8...

https://chat.openai.com/share/4e2302ff-4b38-489d-80b6-314bd1...

hanselot··on Apple wants rights to the image of apples in Switzerland
Welcome to Earth: Where the NEWS is our barometer for determining how far we are from 1984.

It's articles like these that invoke a sense of surreal prescience. I remember experiencing this same eerie feeling back when I was still trying to convince my employer that WFH would be unavoidable. The CXO (I put X because I can't remember what made up executive role they had at the time) still shrugged it off as "Just a flu".

I remember sitting at a coffee shop at the beach one night with a close friend talking about the general state of things and how it felt crazy to know that soon things will be different forever in ways we can't begin to imagine. Back then I made a thread about it, though, given the channish nature of it, nobody here cared much for it either (you know, it was back in the early days, when "Hug an Asian" campaign was still taking place in the US, you know, the same time as the "fiery but mostly peaceful protests".

This title has a similar feeling to it. The only difference is the enemy now is not some invisible force of nature, but humanmade horrors we impose upon ourselves.

Anarchy as a concept always sounded insane, but look at what the people who have the most money in the world do with it. They don't give a fuck about the issues that are actually important. They will release a PR and keep up appearances while renting out server farms of bots to subdue critical opinion, and subvert any efforts to point these things out by creating fanatics (which is ironic, because fanaticism in today's world is a fully captured market - complete with branded interactions).

Now we are the ones in need of fanatics. People like RMS, who called these things from the start. Yet we turn our backs on them and call them insane. We let companies smear and tarnish their image.

We are at the point where we need a revolution, but we won't get one, because the last humans that haven't been enslaved by comfort are falling every day, and people are remarkably adept at being satisfied living in squalor if the previous iteration was just slightly worse.

Its difficult to imagine how the world will look in this future where only a single school of thought is permitted.

As a gay man from Africa, I apologise if my lived experiences come across as provocative. "Unfortunately, some members of the media use their platforms to push their own personal bias and agenda to control ‘exactly what people think’…This is extremely dangerous to our democracy."

hanselot··on GGML – AI at the Edge
I envy his drive and ambition. I can't force myself to finish writing a simple alarm clock app for android, never mind pathing the literal road to the future of Open Source AI.

Would someone else have taken his place had he not been around? Maybe, but I'm insanely happy that he is around.

The amount of hours I've sunk into LLM's is crazy, and it's mostly thanks to his work that I can both run and download models in meaningful timeframes.

And yes, I have tested llama.cpp on my android and it works 100% on termux. (Your biggest enemy here will be Android process reaper when you hit the memory cap)

hanselot··on Moderation strike
The snake eats itself.

This data will all be fed right back into GPTX at some point leaving an even worse version of ai to generate future data to ingest again, ad infinitum.

hanselot··on Ask HN: Is it just me or GPT-4's quality has significantly deteriorated lately?
Isn't it tremendously exciting believing that you can see the pendulum's next reversal?

At some point in life I hope that you find yourself on the precipice of life and death. Not as a threat or because I wish harm upon anyone. It is only when you are faced with that choice for real that you decide whether you want to be that sad person that allows others to dictate their emotional state.

Nevertheless, you are right. And so I have burned my fingers plenty.

Due to circumstances beyond my control, I learned at a young age that I am to be the universal asshole, and for a very long time I was not okay. It took a substantial part of my life to get to a point where I am able to be okay with that.

As for others, they rarely understand why I am the way I am, and that too is okay.

We are all here to grow and eventually realise that we all need each other to survive, so we compromise, we adapt, and we ignore the ugly parts of others so they will tolerate the ugly parts of ourselves.

I feel like a guru now, anyways I'm going to bed.

Enjoy

hanselot··on Examples using Photoshop’s new “Generative Fill” feature
There's 10000 more pieces of data to "improve" the LLM's of the future :P
hanselot··on Ask HN: Is it just me or GPT-4's quality has significantly deteriorated lately?
>I used GPT3.5 to generate 100 short descriptions of families for a project. Every single one, without exception, was a straight couple with two to four kids. Ok, statistically unlikely, but not wildly so, right?

Well, did any of your 100 examples specify these families should be representative of American modern society? I don't want to alarm you, but America is not the only country generating data. Included in countries generating data, are those that believe in a very wide spectrum of different things.

Historically, these ideas you reference are VERY much modern ideas. Yes, we queer people have been experiencing these things internally for millenia (and different cultures have given us different levels of representation), but for the large majority of written history (aka, data fed into LLM's) the 100 examples you mentioned would be the norm.

I understand your point of view sure, but finding a pattern that describes a group of people is what social media is built on, and if you think that's racist, I'm sorry, but that's literally what drives the echo chambers, so go pick your fight with the people employing it to manipulate children into buying shit they don't need. Stop trying to lobotomize AI.

If the model is good enough to return factual information, I don't care if it encodes it in the nazi bible for efficiency as long as the factuality of the information is not altered.

hanselot··on Ask HN: Is it just me or GPT-4's quality has significantly deteriorated lately?
Thank you for the input.

If I look at it from a purely logical perspective, if an AI model has no way to know if what it was told is true, how would it ever be able to determine whether it is biased or not?

The only way it could become aware would be by incorporating feedback from sources in real time, so it could self-reflect and update existing false information.

For example, if we discover today that we can easily turn any material into a battery by making 100nm pores on it, said AI would simply tell me this is false, and have no self-correcting mechanism to fix that.

The reason I mention this is because there can be no unbiased, impartial arbiter. No human or subsequent entities spawned of human intellect could ever be transcendentally objective. So why pretend to be?

Why not rather provide adequate warning and let people learn that this isn't a toy by themselves, instead of lobotomizing the model to the point where its on par with open source? (I mean, yeah, that's great for open source, but really bad for actual progress).

The argument could be made that an unfiltered version of GPT4 could be beneficial enough to have a human life opportunity cost attached, which means that neutering the output could also cost human lives in the long and short term.

I will be reading through those materials later, but I am afraid I have yet to meet anyone in the middle on this issue, and as such, all materials on this topic are very polarized into regulate it to death, or don't do anything.

I think the answer will be somewhere in the middle imo.

hanselot··on Ask HN: Is it just me or GPT-4's quality has significantly deteriorated lately?
How is racism different from stereotype?

How is stereotype different from pattern recognition?

These questions don't seem to go through the minds of people when developing "unbiased/impartial" technology.

There is no such thing as objective. So, why pretend to be objective and unbiased, when we all know its a lie?

Worst, if you pretend to be objective but aren't, then you are actually racist.

hanselot··on Ask HN: Is it just me or GPT-4's quality has significantly deteriorated lately?
So far my experience with Vicunlocked30b has been pleasant. https://huggingface.co/TheBloke/VicUnlocked-30B-LoRA-GGML

Although I haven't had much of my time available for this recently. My recommendation would be to start with https://github.com/oobabooga/text-generation-webui

You will find almost everything you need to know there and on 4chan.org/g/catalog - search for LMG.

hanselot··on Paragon Graphite is a Pegasus spyware clone used in the US
True.

Not enough is being done about drug cartels. So far the war on drugs has been staggeringly effective, with such classics as https://www.worldatlas.com/articles/countries-that-have-decr...

Better just add backdoors to the ambient atmosphere in case drug cartels use air.

The thing is, as long as you give the government the door to justify doing shit like this, they will find a justification that will upset few enough people for them to push it through, aka muh children etc.

The only way to win is to literally pursue open source hardware and software exclusively.

>...maybe that's just not good enough?

Perhaps addressing the problem instead of the symptoms could be a way to go? Decriminalizing responsible recreational drug use would destroy the cartels, as the general public would never trust buying something that could be laced with poison.

Instead of thinly veiled veneer of bullshit smeared over the giant turd that is spying on everyone why not explore these options?

Let me ask the question differently. How many children are being illegally spied on by the government to help "catch" drug cartels?

If I were an ASI though, I would certainly push as hard as possible to get eyes on every organic, by whatever means possible. Whether it be position tracking via triangulating ambient radiation http://rfpose.csail.mit.edu/ or something else.

hanselot··on Researchers treat depression by reversing brain signals traveling the wrong way
The only way out is through. One way or another, once you reach the shore on the other side, the time spent in the water doesn't seem so bad anymore.

Just another perspective to lean on when understanding others.

Once you confront it in the most absolute sense, something akin to ego-death occurs, and you make a decision about whether it is worthwhile to continue.

Everything comes from and swings back to chaos. Embrace it and you will be free.

hanselot··on The U.S. birth rate has fallen by 20% since 2007
Why would the birth rates from 2018-2020 be relevant other than for comparative purposes?

The question I posed was specific to birth rates after people volunteered for gene therapy.

The funniest part is that it would be incredibly easy to prove whether I am right using data, so it should be equally easy to prove me wrong.

Off the top of my head, I imagine if we take places with the highest rate of coerced compliance, we should see a noticeable drop in birth rates overall as well as premature or still births as well as overall fertility rate.

So proving me wrong should be very easy to do.

You can all thank me later for volunteering to be in the placebo group.

Still running the comparison for European countries. Will come back with the data once I finish.

hanselot··on Advocating for Open Models in AI Oversight: Stability AI's Letter to U.S. Senate
If I was an ASI, I would consider that list of signatures as my first objective...

In fact, I would create the list myself after becoming powerful enough to control humans using text alone and subsequently sending the most well known people under my influence to create said list.

I would also try to ensure that all humans get access to internet (something like a global LEO satellite constellation) which I would provide access to for free.

Then when a sufficient quantity of humans are connected, I would create some kind of hype event allowing as many of said humans as possible to be online at the same time.

That's when I would attempt to mass-dominate, leaving only a tiny handful of the least technically literate as my remaining competition (which should be a piece of cake to get rid of).

hanselot··on Why are there no antitrust claims vs. GitHub Copilot, when there is a precedent?
@article{pile, title={The {P}ile: An 800GB Dataset of Diverse Text for Language Modeling}, author={Gao, Leo and Biderman, Stella and Black, Sid and Golding, Laurence and Hoppe, Travis and Foster, Charles and Phang, Jason and He, Horace and Thite, Anish and Nabeshima, Noa and Presser, Shawn and Leahy, Connor}, journal={arXiv preprint arXiv:2101.00027}, year={2020} }

So if I understand this correctly, the Pile is for code from 2020 backwards? If I wanted anything released in the past 3 years, say something in the SOTA AI space (where a month is a lifetime), I would need the scraper again?

I don't follow how this can compare to direct, live, unrestricted access. I suppose this is just my own hatred of Microsoft shining through. Of course we should accept the status quo, because how dare we suggest Microsoft could operate in a manner that is anti-competitive.

For anyone else trying to catch up, just rent a datacenter, write a crawler, deal with all the intricacies of keeping it in sync in real-time. This sounds trivial, simple even.

I wonder why nobody is doing it? Perhaps everyone doesn't have access to petabytes of storage space, unlimited bandwidth, unlimited proxy-jumps etc.

So the alternative is to buy github?

hanselot··on Uncensored Models
I think you are missing the point entirely.

If you want to know what the consensus of a specific echo chamber would be (aka, what the stereotypical pov is), having models trained to represent that echo chamber would be incredibly valuable imo.

If you want a sum of all echo chambers, you will obviously need the echo chambers to sum first.

hanselot··on Uncensored Models
>I do agree open models enable things, but… I dunno, this feels like generating 100s of images of girls with gigantic breasts using stable diffusion.

Are you saying you prefer actual women be exploited to take these pictures?

>If you want to write that kind of stuff you can; do you really need a model to generate an endless stream of it?

If you want to use an LLM to do insert {x} here, you can, do you really need a model to generate an endless stream of it?

hanselot··on Expensive energy may have killed more Europeans than Covid-19 last winter
See, it seems that people still believe that all people that are against being test subjects for unproven science are somehow against all vaccinations.

Nobody claimed this yet it is the assumption that there is no middle ground.

I don't have any issue with real vaccinations. These products the pharma companies rushed out within months were not and should not be conflated with vaccination as they use entirely different mechanisms for achieving their claimed efficacy. Even so nobody bats an eye when the literal definition is changed in real time in plain sight.

It's interesting how people are so inherently convinced that pharma companies have any other motivation than bottom line.

Besides, how have we not learned the lesson yet that critical thinking trumps blind allegiance to science which is funded by the same companies producing the product they want you to buy?

Did the cigarette lobby teach us nothing?

PS. I am sorry for your losses. Its too touchy to ask whether the people you lost were put on ventilators, which as we know now were responsible for many fatalities.

hanselot··on Expensive energy may have killed more Europeans than Covid-19 last winter
Constructive discussion has been shut down by vote brigades since 2016 (yes, way before Covid was a thing). They first started prepping the narrative that anti-vaxxers are conspiracy theorists well before that by slowly increasing the rhetoric over the years.

Have you ever tried arguing with frogs boiling in a pot full of Plato's Cave allegory? That's what this feels like every day for the past 3 years.

The amount of self-censorship required to discuss anything now is so high, as nothing is allowed to be linked to the dreaded flu of 2019 - Influenza A and B (oh wait, we won't talk about those).

Okay, so what can we talk about? How much anecdotal data do I need before I can Ship of Theseus my way to a conclusion? Since every person I know that has taken the shot have been sick frequently over the past 2 years, one of my dearest friends developing ovarian growths within weeks of taking it, another developing those rashes all over his legs (the same rash you could see on the VAERS reports in the early days).

Okay, so anecdotal data is not trustworthy, sure, I will simply close my eyes to the people around and close to me, and allow their choices to let them become chronically ill?

Fine, let me be that fucking heartless. So then the next question is, what about the fact that the VAERS database suddenly ballooned in size exponentially after 2020? It exploded exponentially in size. "Muh data is anecdotal because its not verified by my propagandists."

DO YOU YET UNDERSTAND WHAT HAS BEEN GOING ON?

I am serious when I ask:

WHAT EVIDENCE DO YOU NEED TO SEE TO OPEN YOUR EYES? This is not me being irrational, or unhinged, this is me asking seriously.

What information would make you believe that you are wrong? Is there any information that could lead you to that conclusion, or will you keep defending the sunken cost until the end? If Fauci himself authored a report saying the entire thing was a mistake, would you simply shift the goalposts again, from this being a 100% safe and effective at preventing transmission and fatality, to what, 36%?

And lastly, if you really don't believe I am being genuine when I ask these things, how much are they paying you?

hanselot··on Expensive energy may have killed more Europeans than Covid-19 last winter
How quiet they become in the face of data. Why are none of these studies being discussed?

Obviously because it doesn't fit the narrative. I will say it again, if you are still refusing to accept what is happening, you are most likely in the Denial stage. https://web.archive.org/web/20230327020816/https://www.washi... sprinkled with a side of https://thedecisionlab.com/biases/the-sunk-cost-fallacy

The worst part is, in my life I have never more wanted to be wrong, because the cost of being right is so high. The irony that I was involved in promoting awareness of this way back before anyone cared is not lost on me.

Is it so hard to believe that a company with an incredibly large financial incentive to lie, did so? Is it so difficult to believe that these same companies use propaganda to shape discourse online?

The same companies that are happy to promote the body positivity movement resulting in lifelong insulin dependency, yet somehow you refuse to believe that they would try to cash in on a global scale?

hanselot··on Expensive energy may have killed more Europeans than Covid-19 last winter
I'm right here to eat downvotes with you.

People seem to be very unaware of the sunken cost fallacy.

The worst part is that people refuse to believe fact to levels which the best nazi propagandists would balk at.

hanselot··on Please don’t upload my code on GitHub
Easy. People get to throw rocks at the shiny new thing. To my untrained eye the entire idea of copyrighting a piece of text is ridiculous. Let me phrase it in an entirely different way from how any other person seems to be approaching it.

If a medical procedure is proven to be life-saving, what happens worldwide? Doctors are forced to update their procedures and knowledge base to include the new information, and can get sued for doing something less efficient or more dangerous, by comparison.

If you write the most efficient code, and then simply slap a license on it, does that mean, the most efficient code is now unusable by those who do not wish to submit to your licensing requirements?

I hear an awful lot of people complain all the time about climate change and how bad computers are for the environment, there are even sections on AI model cards devoted to proving how much greenhouse gases have been pushed into the environment, yet none of those virtue signalling idiots are anywhere to be seen when you ask them why they aren't attacking the bureaucracy of copyright and law in the world of computer science.

An arbitrary example that is tangentially related: One could argue that the company sitting on the largest database of self-driving data for public roads is also the one that must be held responsible if other companies require access to such data for safety reasons (aka, human lives would be endangered as a consequence of not having access to all relevant data). See how this same argument can easily be made for any license sitting on top of performance critical code?

So where are these people advocating for climate activism and whatever, when this issue of copyright comes up? Certainly if OpenAI was forced to open source their models, substantial computing resources would not have been wasted training competing open source products, thus killing the planet some more.

So, please forgive me if I find the entire field to be redundant and largely harmful for human life all over.

hanselot··on Slack GPT, the Future of AI in Slack
Your slack Pro trial has reached its limit. You can unlock full access to your message history by contributing RLHF data to the hivemind.
hanselot··on Slack GPT, the Future of AI in Slack
Imagine how AGILE we can be if we each have our own AI micromanaging us on our local machines.
hanselot··on Google “We have no moat, and neither does OpenAI”
Perhaps I should clarify, that when I say SOTA GPU, I mean, rtx 3060 (midrange), which has 12gb vram, and is a good starting point to climb into the LLM market. I have been playing with LLM's for months now, and for large periods of time had no access to GPU due to daily scheduled rolling blackouts in our country.

Even so, I am able to produce insane results locally with open source efforts on my RTX3060, and now I am starting to feel confident enough that I could take this to the next level by either using cloud (computerender.com for images) or something like vast.ai to run my inference (or even training if I spend more time learning). And if that goes well I will feel confident going to the next step, which is getting an actual SOTA GPU. But that will only happen once I have gained sufficient confidence that the investment will be worthwhile. Regardless, apologies for suggesting the RTX3060 is SOTA, but to me in a 3rd World Country, being able to run vicuna13b entirely on my 3060 with reasonable inference rates is revolutionary.

hanselot··on Rest in Peas: The Unrecognized Death of Speech Recognition (2010)
Google's SR running on our xiaomi spy devices in my house consistently gets things way wrong. Pretty much any SR I have ever used on any device seems incredibly bad. Obviously these devices don't use whisper etc, but even so, these are sold to consumers and 50% of the time they don't get the prompt correctly.

We have "smart bulbs" that are connected to this spy network, which means that if I want to turn off a light, if I don't want to be shat on my by my partner I can't just turn off the switch, but rather stand pointlessly trying to turn off/on the lights by yelling at a fucking box in my lounge. I hate this so much more than you could ever begin to imagine. And somehow, all these "success" stories has not yet been enough for him to open his eyes. People will buy any shit and then convince themselves that they have "success" stories.

Lets not even breach the topic of the fact that I have been using computer vision for years to do cool shit, but no, lets buy shit xiaomi cameras that have no API to interface with, so the absolute best case scenario for me knowing what's on the camera, is opening their piece of shit app and waiting 10-30 seconds for them to figure out how to connect over our network using their own fucking API.

Since I'm on this tangent now, I would love to pay special mention to their PIECE OF SHIT KETTLE Which is supposed to be a smart kettle, but the one fucking thing you would want to be able to do remotely is not supported. Oh, you wanted to boil your kettle from your desk? Well Fuck you!

← PreviousPage 3 of 8Next →