@self_made_human's banner p

self_made_human

amaratvaṃ prāpnuhi, athavā yatamāno mṛtyum āpnuhi

16 followers   follows 0 users  
joined 2022 September 05 05:31:00 UTC

I'm a transhumanist doctor. In a better world, I wouldn't need to add that as a qualifier to plain old "doctor". It would be taken as granted for someone in the profession of saving lives.

At any rate, I intend to live forever or die trying. See you at Heat Death!

Friends:

A friend to everyone is a friend to no one.


				

User ID: 454

self_made_human

amaratvaṃ prāpnuhi, athavā yatamāno mṛtyum āpnuhi

16 followers   follows 0 users   joined 2022 September 05 05:31:00 UTC

					

I'm a transhumanist doctor. In a better world, I wouldn't need to add that as a qualifier to plain old "doctor". It would be taken as granted for someone in the profession of saving lives.

At any rate, I intend to live forever or die trying. See you at Heat Death!

Friends:

A friend to everyone is a friend to no one.


					

User ID: 454

Thankfully, we have wildcard rules allowing us to mod people for repeatedly toeing the line, even if they can claim with a straight face no individual line has been crossed. This is rarely used, unless someone has been consistently obnoxious, or low level obnoxious while just avoiding what they perceive as the threshold for mod action.

Regarding Hlynka's second chances (all 25 of them), in general we give additional leeway to longterm users in good standing.

If you have multiple AAQCs, a history of quality posts and are generally in good standing, then if you have an off day, the mods will consider that and go easier on you. A ban might be shortened, or converted into a warning. A warning might be replaced by "knock it off" that doesn't go on the record.

But the leeway extended is not absolute, or infinite. Hlynka pissed me off to no end, beginning with unwarranted personal attacks and casual racism, so if I hadn't recused myself from moderating him, he'd have been shown the door ages ago. But this ban came when his friends in the mod team, people who had worked alongside him for years back when he was a mod, had enough.

I would not recommend Zorba list his business address, it's liable to receive mail bombs.

If some community was daft enough to demand me to perform penance in the form of 2 pages of essay homework, then leaving aside why I would want to come back, that's going straight to ChatGPT.

[User has been banned for this comment]

Cause of ban: Self-identification as an alt of permabanned account

Expiry of ban: Heat Death of the Universe

Ban Appeal Guidelines: Wait till expiry of ban and raise a ticket

Your specific issue hasn't cropped up for me. In my case, it's because I'm on mobile and frequently need to change tabs to hunt down links and citations, and when I return to my comment it's been flushed from memory.

Feel free to report your specific instance as a bug on Github, though I think it's rare enough that I'm not sure it'll be easy to reproduce reliably or fix.

Fair enough, but I doubt Hlynka will ever change his spots, so I strongly expect that if he was to return, he'd be back at it again, and his identity would be obvious.

There's a difference between having an alt, and having an alt after you were permabanned. Off memory, I can't recall an example of the latter that the mods knowingly condone, but it could well be true.

You're correct that this isn't a correctional facility, but you're mistaken if there isn't moderation involved in keeping our tiny enclave in the internet wastes that way.

Anyone who doesn't like it is welcome to leave. There plenty of other forums out there. The mods here still take pains to maintain the Mandate of Heaven approval of the supermajority of the users, since without them, we'd be a glorified day drinking club and Zorba's AWS bill would be significantly less justified, if a lot smaller.

But that's that for Hlynka. Making an alt after being permabanned doesn't work as well as you'd like, especially for such a colorful character as Hlynka. Anyone can tell it's him even behind seven layers of VPNs, and anyway, I think he has more of a martyr complex and feels vindicated rather than desiring to sneak in under a new alias.

Even before he started his beef with me, I was frankly confused by many of his posts that were nominated and accepted as AAQCs.

A lot of it came off as incoherent rambling, he was often incapable of writing clearly to save his life. When his post on "Marvin's Marvelous Minutiae" came out, I had an aneurysm reading it, as I've said elsewhere, it was in the "not even wrong" tier as far as I'm concerned.

But despite being a moderator, I don't handle the AAQCs, and the other mods have far more fondness for the man than I do and saw something in him I don't. That earned him a lot of leeway and n-th chances, and his permaban was when even the people positively inclined towards him got fed up with his antics and inability to improve.

My own mod log has hilarious examples from my own arguments with Hlynka, such as a "ban both the chucklefucks if they do it again" from @Amadan. But unlike Hlynka, I am capable of acknowledging error and not doing the same damn thing over and over again, and I stuck to my promise to studiously never engage with him, since I knew that if I did, he'd piss me off enough into violating forum rules about decorum. Hell I even recuse myself from enforcing mod actions against him when I became a mod, because no matter how much they were justified, the optics would be bad. Caesar's moderati must be above reproach and all that.

The best I can say about him is that his anecdotes about military/PMC life were occasionally interesting, but he annoyed the hell out of me with his pugnaciousness, inability to accept being wrong even when corrected with evidence and backup from people who know what they're saying (I'm a dilettante in the field of ML, I just read things), and general incorrigbility.

At least I know I won't miss him, sad as the loss of a regular can be.

Albinism doesn't have negative effects on IQ, but the lack of melanin can fuck up the eye. And as med school processors love to say, the eye is an extension of the brain, from a developmental and anatomical standpoint.

https://eyewiki.aao.org/Albinism

Iris transillumination: The iris in albinism has little to no pigment to screen out stray light coming into the eye. On slit lamp exam, the examiner may detect speckled or diffuse transillumination defect. This finding, while common with albinism, is not specific as iris transillumination occurs in diseases unrelated to albinism such as pseudoexfoliation, pigment dispersion syndrome, megalocornea, iris atrophy, and Axenfeld-Rieger spectrum [7]. When present in an otherwise normal individual, this finding may indicate carrier status of a hypomelanotic gene mutation. The iris may be translucent and the margin of the crystalline lens may be visible on transillumunation during slit lamp examination.

Pendular nystagmus: Nystagmus refers to rhythmic, involuntary, conjugate eye movement. Affected infants may have large amplitude with low frequency pattern of eye movement starting at 2-3 months of age, later changing to a pendular form without distinct fast or slow phases. Eye muscle surgery may be considered to reduce nystagmus.

Foveal hypoplasia (absence of a foveal pit): In albinism, the retina does not develop normally before birth and in infancy because of inappropriate retinal pigment epithelium (RPE) pigmentation that is required for macular development. Optical coherence tomography (OCT) can demonstrate an absence of the foveal pit and the loss of normal thinning of the retina. Also, the foveal avascular zone is small or nonexistent with vessels crossing the area 2 disc diameter temporal to the optic disc margin. Foveal hypoplasia is the single most important contributor to poor vision in albino patients. [7]

Abnormal decussation of the visual pathways: Normally about half of optic nerve fibers from each eye decussates at the optic chiasm to the contralateral side, contributing to stereopsis. Albinism is associated with fiber over-decussation, resulting in crossing of up to 90% of fibers to the contralateral side and thus strabismus and loss of stereopsis. Evidence of this abnormality can be detected by 3-lead visual evoked potential (VEP) for proper counseling regarding visual potential of a patient.

Strabismus: Misalignment of the eyes and related anomalous head tilt can occur in association with albinism. Kumar et al reported that the strabismus is seen in higher proportions of those with albinism compared to those with idiopathic infantile nystagmus, suggesting different mechanisms underlying the cause of strabismus in the two disorders. [8] Photophobia: Sensitivity to bright light and glare can occur due to scattering of light within the eye. Patients may prefer to wear sunglasses to reduce their sensitivity to light. Refractive Errors: Both myopia and hyperopia can occur, and astigmatism is very common. Poor vision: Vision can range from normal for those minimally affected to legal blindness or worse (vision less than 20/200) for those with more severe forms of albinism. Near vision is often better than distance vision. Generally, those who have the least amount of pigment (i.e. most severely affected) have the poorest vision.

What I'm getting at is that if you modify phenotypes too much, then you can get subtle knock-on effects, and in this case, you don't have to go full albinism. But we have examples of happy, healthy and functional people with very high IQs, so there are enormous gains to be made with conservative approaches before you start to make painful tradeoffs (like the recent reports of a family in Ireland with a mutation that gives them 20 IQ points extra but causes blindness in their 20s).

You can make the average person much smarter and not have negative consequences.

Well I certainly didn't, if only because I speak better English than he does and that requires some working memory.

I think the specific claim in question was I was unable to parse his clear language and misunderstanding him because it's difficult for me to translate from English to "Indian".

The audacity, I find it even hard to be mad. I did find his takes on machine learning and LLMs to be far worse for my blood pressure, as they were the kind of "not even wrong" mixed with gish-galloping and a pugnacious inability to know when he was conclusively and incontrovertibly mistaken, in a manner not easier to falsify than reading a comment I write, in English.

Anyway, I for one am not sad he's gone, though I have repeatedly had to recuse myself from taking mod action against him however justified it was because of the obvious conflict of interest (and so do other mods with people they have some kind of personal tiff with, unless they're the kind of asshole who has beef with all of us) . I knew this would happen, he can't help himself.

What exactly are you trying to convey through this comment barring pure antagonism? Consider this a warning. I don't extend a short ban only because you don't have any prior offenses in the mod queue, but personal attacks of this nature, especially so non-sequitur, are not kosher.

Edit: After this received a bunch more reports, I'm upgrading you to a day's ban. I feel like even a regular in good standing would deserve rebuke for an uncalled for and random attack of that nature, and you haven't earned that. Do not do that again.

Not really. I haven't run into any task that Opus could perform but GPT-4 couldn't, at least in a reasonable amount of testing. And the latter is more flexible IMO, at least in terms of content guidelines, even if neither is perfect on those grounds.

ChatGPT paid certainly has more perks like DALLE and customs GPTs and plug-ins , depending on how you value those. But the core product, in the form of the intelligence and competency of the AI assistant, is much the same so I wouldn't change if one wasn't significantly cheaper.

Which, in the Kendi framework, makes you a racist. HBD-blogger Jayman is clear on this too. All your genetic engineering, polygenic screening, embryo selection, even if fully voluntary, is still "Nazi shit."

I'm exceptionally fortunate that I couldn't care less what Kendi thinks, and his stream of thought can't really hurt me.

After all, Scott Alexander endorses genetic engineering and polygenic screening, and look at that, he makes 2 orders of magnitude more money than I do and holds an AMA registration. What an awkward situation, given that he has a far larger audience and more content than yours truly. Hardly something being crushed like you claim.

If that makes me a Kendian-racist, my question is, so? He makes a lot of things racist. His success and reach is not infinite.

I think humans live better lives than cats, ants, donkeys or horses. And it's not all thanks to opposable thumbs. And ants certainly commit infanticide. That's half the fun for them when it comes to raiding other colonies, even of the same species, the other half being enslaving the babies I guess.

No, intelligence is not what produces conflict. Dumb creatures have plenty of conflict. Likely more conflict, if one looks at the rates of violence and murder in a chimpanzee tribe and compares it to the average human one. We live better lives.

Your post is a potent illustration of a point I have made many times that fell upon deaf ears.

HBD is a fact.

Denying it is an erroneous rejection of reality. It's the thing that doesn't go away when you stop believing in it.

The implications, however, are a matter of preference, and further downstream are the policies you wish to enact.

Looking at the same disparity in group IQs, one can:

  1. Use that as a justification for racism, in the strong sense, like disbarring them from the commons. Kicking them out of a country. Banning miscegenation.

  2. Still discriminate, but on more utilitarian grounds, such as by demanding IQ tests of prospective immigrants so that the baseline level of IQ necessary for a wealthy, comfortable and cohesive society does not become diluted into dysfunction. Are you black but still above 100 IQ? You're welcome. It's regrettable that this disqualifies 70% of your compatriots, but it's not personal.

  3. Argue that this justifies more redistribution. After all, we aid the sick and pay for the disabled. Why not prop up those who are cursed to be dumber because of their genes and had no say in the matter?

  4. Or like me, argue that this is a problem to be solved, by means such as genetic engineering and polygenic screening. We know thousands of genes that all weakly contribute to intelligence. We can perform embryo selection, and with iterated embryo selection get nigh arbitrary gains limited only by our knowledge of causative genes and number of cycles we wish to churn through. This will almost certainly pay for itself. There is no fundamental biological reason that skin color should have any correlation with ones intelligence, anymore than painting an F1 car a different shade will change lap times beyond measurement error. I strongly suspect that even facial physiognomy can be conserved as long as you leave enough damn room in the skull for the brains, to head off claims that people have made before. Hooked noses do not make Jews smarter, flatter ones do not make Africans dumber. Melanin production has no reason to affect neuronal function, unless you really screw the pooch and end up with albinism.

I obviously endorse 4. Intelligence is as close to an unalloyed good as it gets. I do not think 1 is a good idea.

2? I can get behind it, at least until root-cause modification can be enacted. Most sane countries have this to a degree, if they privilege educational attainment, which is a decent proxy for intelligence. Not enough, sadly, and they ruin it by letting in "asylum seekers", keeping illegals and so on.

3? Not a fan. Eventually we'll all be economically obsolete, so I certainly would prefer UBI that does not care that I'm OOMs dumber than a hypothetical AGI. But I will tolerate no more redistribution, while humans have to pull their weight, than is absolutely necessary to ensure that people don't starve to death, have a permanent domicile, education and medical care. Not additional participation dollars for being too incompetent, beyond what anyone should receive.

But what remains obvious, for those with eyes to see, is that group differences in intelligence exist, have enormous empirical impact, that denying this inevitably leads to suspicions of systemic racism that's keeping otherwise fine folk down, and all the consequences of a gaping ideological blindspot in one's understanding of the world.

Yes. The maximum bound on the "badness" of vaping is about 5% of that of cigarettes for equivalent use, IIRC. That is the maximum, there is evidence the net impact is close to nil.

Suffice to say I did read research papers before I decided to take it up (from being an adamant non-smoker) because I was convinced that barring nicotine addiction, there are no significant downsides. Nicotine isn't even notably bad for you, closer to caffeine.

I am perfectly happy with being a productive member of society. I like helping people and being useful.

Unfortunately, I expect that humans will be obsolete in that regard, so you bet I'm willing to be content with mild hedonism and doing whatever I like. Nothing that fundamentally warps my psyche or is of the level of addictiveness as opioids or wireheading, but video games? The ones I like are abstracted away enough from Skinner Boxes that I make the unprincipled exception of being willing to play them forever.

There are better meds. Ketamine is promising for treatment resistant depression. There's transcranial magnetic stimulation and electroconvulsive therapy.

If your meds aren't working (well, maybe they are, but you certainly don't seem to be not depressed), then alternatives exist. I look forward to getting ketamine treatment when I move somewhere where that's an option.

You're not teaching the average person Quantum Mechanics. You're not teaching someone with an IQ of 80 how to code or calculus, at least not beyond rote memorization.

My experience with dumb people (as patients) is that they have a worse time keeping track of multiple, interacting bits of information, an inability to predict what seems like obvious interactions and consequences, and worse memory in general.

When it comes to moral cognition, that obviously makes them worse consequentialists, and to the extent that Deontology is unworkable without a mechanism for breaking ties, bad at choosing what to sacrifice. They are also less likely to notice positive sum opportunities and defect, and while more intelligence does offer more opportunities to get a fast one over other people, it also usually makes one notice that life is an iterated game and that in most cases that's not the best idea.

I am skeptical about the "how would you feel if you if didn't have breakfast" meme, at least for 80 IQ people. Lower than that? Almost certainly. But 80 IQ people are slow but largely functional and capable of independent operation in complex society, and can probably manage that, most of the time.

It is worth noting that there are functional complex societies where the average IQ is below 80 (I live in one). In a society where the average IQ is 100, having an IQ of 60 or 70 is more likely to be due to gross cognitive issues manifesting as retardation. Whereas in a society where that is the norm, they're still functional, just dumber. In other words, something deeply broken versus something that is working as intended, just subpar in comparison.* There can also be enclaves of higher IQ people capable of running the place even if the masses are dumb (India is incredibly heterogenous due to millenia of endogamy. Brahmins are almost certainly smarter. TamBrams are notoriously so. Modeling it assuming a normal distribution of intelligence here would be a gross error)

*Racism not intended, but that is the difference between a healthy chimp and and a very retarded human. They might perform similarly on IQ tests, but one is a functional animal operating comfortably in its niche, the other isn't.

Helldivers 2 added mechs, after a long and grueling campaign where players had to fight in the hundreds of thousands to liberate the planet producing them.

They're great. Big stompy and utilitarian tin-cans that pack 1000 rounds of 7.62x51 Fuck You and 12 rockets of fuck you in particular. I'm told they're similar to Mantis mechs from Halo in terms of design, but I'm a 40k fan so I see Dreadnought.

Unfortunately, much like the Dreadnought, they can quite easily turn into a coffin if anything shoots at you.

Bugs, thankfully, do not have rocket launchers or high caliber projectiles, so the mechs handle them with ease. The biggest baddies in the game go down with 3 good rockets to the face. If you can't chew through a hundred bugs with a thousand rounds, git gud.

The robots? GG. They have more missiles than once covered the Fulda Gap and don't believe in MAD. Enjoy being blown up I guess.

Still great fun, but I do wish they were a tad bit tankier. They can blow up from a stiff breeze or stepping on flowers (explosive flowers, but they don't kill armored humans), friendly fire or just turning so fast the hitbox of the missile intersects with you before it leaves the tube.

If I'm effort-posting, I regularly save all the text to my clipboard or use a notes app.

I'd love to have drafts auto-saved, and it's something that's been suggested to Zorba, but he's got enough on his plate as is.

I would wager it's because most old games you would need to emulate are de facto abandonware. The studios might have dissolved, who holds the IP might be unclear, and even companies like Nintendo are often lazy about making their cold cuts available to customers.

As opposed to music or movies, at least examples that aren't ridiculously niche.

And emulators are not necessarily illegal, endless DMCA litigation and baseless lawfare aside. A clean-room implementation would be entirely above board, and some of the more popular emulators, especially for more recent platforms, take pains to keep it that way.

Far be it from me to cast moral judgement on the use of massive amounts of matrix multiplication.

Yeah, a 6 gb card means you're fucked unless you want to run baby models, maybe a 4 bit quantized 3b or 7b model I guess.

I honestly don't know why companies are so allergic to people using their models for erotica, even if that's not my cup of tea. But I remember AI Dungeon fondly before they fell out with OAI and then became trash, and I don't think they've gotten better.

With million and ten million token context windows, the opportunity to begin an interactive story that never ends still tantalizes me. I don't want to have to write my novel (though I do like writing). I just want to lay out my ideas and have the grunt work taken care of. It'll happen, sooner rather than later.

The levels of horny on main are remarkable. Have you tried using Mixtral instead? Assuming you have the hardware to self-hosted, I'm sure there finetunes with the limiters removed. It's a 3.5 tier model, or at least the released MOE version is, but I'd presume that's good enough for the man-made horrors beyond my comprehension you're trying to engage in haha.

I've had a reasonable amount of hands on time with Claude Opus, and I would rate it as indistinguishable from GPT-4 in terms of usefulness, or at least I can't tell any obvious disparities after tens of thousands of tokens of conversation.

It is still however, more of a prude, at least as compared to GPT-4. I asked it to try extending an excerpt from a chapter from my novel, and it categorically refused because it included scenes of "body horror". GPT-4 didn't care.*

This is an improvement over Claude 2, or 2.1, but only because those were beaten into nigh uselessness.

(I'm ever curious as to when an AI can write fiction in my style as well as I can, because then I'm going to fucking retire, but that day is not today. Best guess is 1-2 years.)

In other words, I think Claude 3 isn't really a breakthrough (barring the long context window, and no degradation in quality with it according to needle in the haystack assessment, if you need it), but at least there is a legitimate GPT-4 competitor out there. I'd love to evaluate Gemini Ultra myself, but I'm not going to pay for it, but so far GPT-4 being about two years old if counting from completion of initial training, suggests that OAI still has a massive lead and they'd be retarded if they squandered it.

What excites me more than Clod (a less deserved nickname now, admittedly) or Gemini themselves is that they'll likely prompt OAI to launch the next GPT, be it something called 4.5 or an outright 5. Or at least do it sooner instead of milking 4 I guess.

Edit: Mf hasn't even heard of my book, though it should be within the knowledge cutoff in late 2023. Guess they're not scraping Royal Road or it's too niche to remember.

*I've evaluated it in other ways, and once again my takeaway is "GPT-4 parity". That includes applied Bayesian Reasoning and arcane questions about orbital mechanics and megastructures. I had both GPT-4 and Claude Opus crosscheck each other, and they both concluded the other was doing a good job.