I’m a software developer by training with an interest in genetics. I am currently doing independent research on gene therapy with an emphasis on intelligence enhancement.
GeneSmith
Karma: 3,498
GeneSmith’s Shortform
Billionaires read LessWrong. I have personally had two reach out to me after a viral blog post I made back in December of last year.
The way this works is almost always that someone the billionaire knows will send them an interesting post and they will read it.
Several of the people I’ve mentioned this to seemed surprised by it, so I thought it might be valuable information for others.
Can you link to a source about the increased cancer risk? Every source I’ve seen on this subject indicates retinoids DECREASE cancer risk, though I agree with you that this doesn’t make much intuitive sense.
According to Claude they were first studied for cancer but the first actual FDA approval was for acne.
I’ve read the article and notice that all the negative side-effects listed are due to issues with oral Retinoids rather than topical Retinoids.
Several of the articles I’ve read indicate that topical retinoids actually DECREASE the risk of cancer, though I agree this is confusing since they supposedly increase cell turnover rates, which should theoretically cause more mitosis-induced mutations to occur. But Retinoids are actually frequently used as anti-cancer drugs.
Google’s AI summarizer says you the mechanism of action is promotion of cell differentiation and inhibiting the progression of pre-malignant cells to malignant cells.
It also reduces “keratinocyte differentiation and decrease keratin deposition” which makes skin more fragile, and it increases sensitivity to UV light.
This is correct of course and why I specifically recommend in the post that people wear sunscreen when using retinoids.
Hard to say. Retinols are recommended as a preventative. Though incidentally I have noticed acne fading much faster after using them. I suspect this is because they speed up the healing process of the skin.
Why you should be using a retinoid
I’ve started a gene therapy company, raised money, opened a lab, hired the inventor of one of the best multiplex gene editing techniques to be our chief scientific officer, and am currently working on cell culture experiments with the help of a small team.
I may write a post about what’s happened at some point. But things are moving.
Can you explain in more detail what the problems are?
You can definitely extrapolate out of distribution on tests where the baseline is human performance. We do this with chess ELO ratings all the time.
I don’t think this is the case. You can make a corn plant with more protein than any other corn plant, and using standard deviatios to describe it will still be useful.
Granted, you may need a new IQ test to capture just how much smarter these new people are, but that’s different than saying they’re all the same.
Apart from coming across as quite repulsive to most people, I wonder at the cost and success rate of maturing immature oocytes.
This is already an issue for child cancer patients who want to preserve future fertility but haven’t hit puberty yet. As of 2021, there were only ~150 births worldwide using this technique.
The costs are also going to be a major issue here. Gain scales with sqrt(ln(number of embryos)). But cost per embryo produced and tested scales almost linearly. So the cost per IQ point is going to scale at like , which is hilariously, absurdly bad.
I’d also want to see data on the number of immature oocytes we could actually extract with this technique, the rate at which those immature oocytes could be converted into mature oocytes, and the cost per oocyte.
So a human with IQ 300 is probably about the same as IQ 250 or IQ 1000 or IQ 10,000, i.e. at the upper limit of that range.
I would be quite surprised if this were true. We should expect scaling laws for brain volume alone to continue well beyond the current human range, and brain volume only explains about 10% of the variance in intelligence.
Without any double-strand breaks, base editors are less toxic to cells and less prone to off-target effects.
It’s worth noting that most base editors actually DO involve nicking of one strand, which is done after the chemical base alternation to bias the cell towards repairing the non-edited strand.
The editing efficiency of non-nicking base editors is significantly lower than that of nicking versions (though the precise ratio varies depending on the specific edit site)
Finally the cell’s enzymes also notice a mismatch between the strand with the new template DNA and the old strand without it, and decide that the longer, newer strand is “correct” and connect it back to the main DNA sequence.
It’s worth noting that this only happens some of the time. Often the cell will either fail to ligate the edited strand back together or it will remove the edited bases, undoing the first half of the edit.
In regards to bridge RNAs, I do not yet believe they will work for any human applications. The work in Hsu’s paper was all done in prokaryotes. If this tool worked in plants or animals, they would have shown it.
In fact, despite what human geneticists often say about epistasis being minor and rare, plant genetics people seem to find that interactions between genes are a big deal, explaining most (!) of the variance in crop yields.[5] So, if I’m not mistaken, “of all these genetic variants statistically associated with the polygenic trait, what’s the best subset of edits to make, if I want the largest expected impact” is a nontrivial question.[6]
I was curious about the finding of epistatic effects explaining more of the variance than traditionally assumed, so I took a look at the study you referenced and found something worth mentioning.
The study is ludicrously underpowered to detect anything like what they’re trying to show. They only have 413 genetically distinct rice plants in the study, compared to 36,901 SNPs.
This study is underpowered to detect even SNPs that have an effect on the trait in question, let alone epistatic effects. So I don’t give their results that much weight.
I agree with your overall conclusion though; I think we’ll likely see the first applications of polygenic embryo selection in animals (perhaps cows?) before we see it in humans.
I have to say, I don’t find these bullet point lists developed by AIs to be very insightful. They are too long with too few insights.
It’s also worth pointing out that the AI almost certainly doesn’t know about recent developments in iterated meiosis or super-SOX
If we actually even tried to launch a global scale effort to genetically engineer “superhumans” it might take at least 10 years to develop the technology
This is definitely wrong. A global effort to develop this tech could easily bring it to fruition in a couple of years. As it is, I think there’s maybe a 50% chance we get something working within 3-5 years (though we would still have to wait >15 years for the children born with its benefits to grow up)
From the current advances in AI it does not seem plausible that we have 30 or more years before foom.
I agree this is unlikely, though there remains a non-trivial possibility of some major AI disaster before AGI that results in a global moratorium on development.
This forum is quite concerned about AI alignment. Aligning superhumans might be much much more difficult than with AI. At least with AI there is known programming—with humans the programming is anything but digital (and often not that logical).
I agree that aligning superhumans could potentially be a concern, but consider the benefits we would have over AGI:
Literally hundreds of years of knowledge about the best ways to raise and train humans (not to mention a log of biologically primed psychology that helps that training process along)
The ability to edit and target other traits besides intelligence such as kindness or altruism
We have guaranteed slow takeoff because they will take over a decade to reach cognitive capabilities of the level of our smartest people.
Consider also the limitations of even genetically engineered superhumans:
They have no ability to rapidly and recursively self-improve
They cannot copy themselves to other bodies
While these don’t guarantee success, they do give us significant advantages over an AGI-based path to superintelligence. You get more than one shot with superhumans. Failure is not catastrophic.
I’ve switched to Claude when Opus came out. If the delta between whatever OpenAI has and the next best model from a more ethical company is small enough then it seems worth it to me to switch.
When Opus came out and started scoring about as well on most benchmarks I decided the penalty was small enough.
Send me an email: genesmithlesswrong@gmail.com
Give me a summary of your background, what stage of your career you’re at, and your motivations for working on this tech.
Send me an email: genesmithlesswrong@gmail.com with some info about what your skills are. There are some companies you might be able to apply to if you’re at working age. If you’re in university you can upskill in certain areas.
I appreciate the offer. It’s going to be a long time before we start human trials. At least five years, probably longer.