Examples of AI Increasing AI Progress https://www.alignmentforum.org/posts/W3tZacTRt4koHyxbr/examples-of-ai-increasing-ai-progress
Recent language model results from DeepMind https://www.lesswrong.com/posts/HXDkCtk9tae5wFmjG/an-173-recent-language-model-results-from-deepmind
DeepMind: The Quest to Solve Intelligence https://youtu.be/kFlLzFuslfQ
Tiny Motors Take a Big Step Forward: First-Ever Solid-State Optical Nanomotor https://cockrell.utexas.edu/news/archive/9527-tiny-motors-take-a-big-step-forward
Interview with head of research at Waymo and DL pioneer Drago Anguelov https://www.listennotes.com/podcasts/gradient-dissent/drago-anguelov-robustness-Pi_l0QmZcXN/
A Media-Fueled Social Panic Over Unmarked Graves — Not a single body has been unearthed. But Canadians wouldn’t know it from the false information reported in The New York Times. https://quillette.com/2022/07/22/how-fake-news-in-the-new-york-times-led-to-a-canadian-social-panic-over-unmarked-graves/
“Resumes of top PRC leaders are qualitatively different from those of western politicians...most people familiar with the process believe that the CPC *is trying* to promote meritocracy even at the highest levels of government.” https://threadreaderapp.com/thread/1548325100302151683.html
Drone Airspace: A New Global Asset Class https://www.cspicenter.com/p/drone-airspace-a-new-global-asset
“According to a survey, more than 70% of doctors said they would administer antibiotics to treat asymptomatic bacteria found in urine tests, disobeying recommended guidelines” https://scitechdaily.com/over-70-of-physicians-still-prescribe-unsafe-antibiotics-which-can-be-deadly/
"Actually, the Russian Economy Is Imploding" [foreignpolicy.com] https://archive.ph/9yYTy
Comment on Open Borders by Anon https://betonit.substack.com/p/incidence-not-insanity/comment/7888309
This problem is more widespread than you might think.
For example, when archeologists ponder examining a delicate historical artifact they need to take into account that future technological progress will reduce the risk of damaging it.
More generally, it can be argued that it is not worth it to try to solve certain problems at this point because it will become much easier and cheaper in the future when we know more and have better tools.
But you have to start at some point. This is the exploration-exploitation paradox of rapid progress.
Dysgenics (negative correlation between fertility and IQ and increasing mutational load in humans due to a low mortality rate) means we wont be solving future problems more quickly or cheaply. This is explained here -> https://qr.ae/pvMrj3. We need to solve the hard problems now.
"Yes, the human genome is degrading. This is a well-established, noncontroversial finding. This phenomenon is called “increasing mutational load” and is based on concepts developed by one of the great geneticists, H. Muller, roughly 70 years ago.[1]
The theory
Harmful mutations come into being all the time. The average newborn has 50–100 new mutations. Though most are harmless, about 1 to 4 of those are harmful.[2] Normally, natural selection causes people having those mutations to die out or not have children, so eventually, those mutations get eliminated.
But our lives aren’t natural anymore. At the beginning of the 20th century, people having genes predisposing them to diabetes would have died young. That’s the normal process of natural selection. Now, those people get life-saving insulin, so they live normal lives and have as many children as anyone else. (Type 1 diabetes, the kind that can appear before you start having children, is highly heritable.[3]) The children of diabetics inherit the genes that make one susceptible to diabetes, so those genes aren’t being eliminated.
The same thing is happening for many diseases that have a genetic component. People that would have died in the past now live nearly-normal lives, and pass on their genes to the next generation.
It’s scary
One investigator calculated that without natural selection, fitness will decline 1 to 3% per generation, and then went on to write the most frightening paragraph I have ever seen in a biological publication:
Thus, the preceding observations paint a rather stark picture. At least in highly industrialized societies, the impact of deleterious mutations is accumulating on a time scale that is approximately the same as that for scenarios associated with global warming ... Without a reduction in the germline transmission of deleterious mutations, the mean phenotypes of the residents of industrialized nations are likely to be rather different in just two or three centuries, with significant incapacitation at the morphological, physiological, and neurobiological levels.[4]
If you don’t normally read biological publications, this paragraph may seem tame to you, but this is as alarmist as biologists ever get.
Not so scary
If medicine and biology keep advancing, they will always stay ahead of the increasing mutational load. Someday, there will be good treatments for asthma, Crohn’s disease, diabetes, obesity, and other genetic diseases, so it won’t matter if the genes causing them to become common.
But if civilization ever crashes, people with multiple genetic defects might not survive."
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1716299/
https://www.pnas.org/content/107/3/961
https://www.dovepress.com/familial-aggregation-and-heritability-of-type-1-diabetes-mellitus-and--peer-reviewed-article-CLEP
https://www.pnas.org/content/107/3/961
[1] Our load of mutations
[2] Rate, molecular spectrum, and consequences of human mutation
[3] Familial aggregation and heritability of type 1 diabetes mellitus and | CLEP
[4] Rate, molecular spectrum, and consequences of human mutation
Who cares about alignment, we're all going to be dead soon anyways.