2050 Predictions

Inspired by a post by Eric Hoel (which I don't agree with in parts, but it's overblown), along with some fellow bloggers, I decided to lay out some predictions for 2050. In other parts of this post, I'll introduce some meta-tools for monitoring data collection, and then I'll address monitoring itself. If you're expecting my (I hope) at least speculative annihilation before 2022, fear not! They'll find out the worst in a few months.

Here's a list for monitoring others before 2050:

• Slime Mold
• Space} Jehan from Atoms vs Bits
• Space} Rohit from Strange Loop Canon
• Roger's Bacon
• Space} Max from My Super Secret Diary
• Experimental Situation
• The creation of this monitoring was more difficult than it was waiting to be. I started here trying to arrange a fairly clear quantitative monitoring, similar to what I actually did for my own post on monitoring for 2021. However, I quickly realized that this is very difficult. First, even super-forecasters can barely make forecasts for more than one year (for geopolitical events, but in fact it probably applies in the broadest sense) (source). In fact, as a result, it seems almost impossible for me to do quantitative monitoring 28 years into the future. Secondly, my quantitative monitoring often depends on the existence of immutable and reliable data sources. Given the basic gambling rate of the links, almost all the keys I choose will finish pressing in 2050.

Following these rationales in this post, I delayed my quality monitoring, because, quality monitoring aside, you cannot simply check the post fact (see command 6 here for more on this issue). I try to mitigate this case in three ways. First, in each successive (potentially fuzzy) divination, I try to connect similar specific subproblems only once. For example, let us understand my divination on AI soldiers:

Systems of (supposedly) artificial origin of mind are accepted into the US military without literally playing the role of a person. This is not apocalyptic, as for example they see in video clips about massacres, but it leads to oversight (in hindsight) and exacerbates the direction of the sensory removal of civilians from wars and incidents.

Universe

It is difficult to measure the degree of terminal theory, but hopefully the first sentence does not give too much room for interpretation. For example, if it is known that the US military will use a drone controlled by artificial intelligence without human participation in the assassination decision by 2050, this will definitely be Jesus. 。

I had a low estimate of the probability of a catastrophe that dramatically changes the forecast of 2050. Pandemic, which wipes out most of the world's population, development of defective IPS, nuclear war, the collapse of the US government, and the explosion of the Yellowstone's super huge volcanoes. I have no clear opinions on the probability of each event, and I trust other people who can be trusted. However, I believe that all the risks are not for Pascal's betting, but to be clearly avoided. It is important to write about these risks, but this article is too complicated, so I depend on all my 2050 predictions that there is no existing catastrophe. If you want to write a version of this article that predicts all the ways we may die by 2050, you can write this way.

On the other hand, it also acknowledges that there is a high possibility that a positive scenario for social change will occur before 2050. This includes the birth of super intelligent AI and uploading the human brain (recently "digital Human"). My prediction is also a condition that these things do not happen.

In other words, these predictions are not the distribution of the future that I expect, but the modal future that I expect. < SPAN> It is difficult to measure the degree of terminal theory, but hopefully the first sentence does not give too much room for interpretation. For example, if it is known that the US military will use a drone controlled by artificial intelligence without human participation in the assassination decision by 2050, this will definitely be Jesus. 。

The second way to avoid becoming an expert is that I think that predicting something is more likely than not. As mentioned earlier, it's not easy to solve this problem, but at least, at least less than half of what I predicted would not be impaired. This also means I bet on these predictions. Nevertheless, if I quote this post in 2050 and see it as evidence that I have a good achievement, these predictions should not be basically counted on my achievement list. Let's allow to make a proposal.

I had a low estimate of the probability of a catastrophe that dramatically changes the forecast of 2050. Pandemic, which wipes out most of the world's population, development of defective IPS, nuclear war, the collapse of the US government, and the explosion of the Yellowstone's super huge volcanoes. I have no clear opinions on the probability of each event, and I trust other people who can be trusted. However, I believe that all the risks are not for Pascal's betting, but to be clearly avoided. It is important to write about these risks, but this article is too complicated, so I depend on all my 2050 predictions that there is no existing catastrophe. If you want to write a version of this article that predicts all the ways we may die by 2050, you can write this way.

On the other hand, it also acknowledges that there is a high possibility that a positive scenario for social change will occur before 2050. This includes the birth of super intelligent AI and uploading the human brain (recently "digital Human"). My prediction is also a condition that these things do not happen.

Predictions #

AI #

In other words, these predictions are not the distribution of the future that I expect, but the modal future that I expect. It is difficult to measure the degree of terminal theory, but hopefully the first sentence does not give too much room for interpretation. For example, if it is known that the US military will use a drone controlled by artificial intelligence without human participation in the assassination decision by 2050, this will definitely be Jesus. 。

Overall Worldview #

The second way to avoid becoming an expert is that I think that predicting something is more likely than not. As mentioned earlier, it's not easy to solve this problem, but at least, at least less than half of what I predicted would not be impaired. This also means I bet on these predictions. Nevertheless, if I quote this post in 2050 and see it as evidence that I have a good achievement, these predictions should not be basically counted on my achievement list. Let's allow to make a proposal.

1. On the other hand, it also acknowledges that there is a high possibility that a positive scenario for social change will occur before 2050. This includes the birth of super intelligent AI and uploading the human brain (recently "digital Human"). My prediction is also a condition that these things do not happen.
2. Despite the innate propensity of speculation about the topic that I don't understand, I finally decided to monitor the fields that seem to have more information than the average. In other words, we decided to monitor the artificial origin of intelligence, biology (technology), welfare, and technology in general. I also arranged some monitoring about cryptocurrency because I couldn't control it myself. If I often make mistakes in cryptocurrencies than in other fields, that would be a great lesson for me. If you are interested in the topic that I decided not to predict, please refer to the section "what I decided not to predict".
3. In general, my view of AI's progress is based on hig h-quality monitoring. If it turns out that at least one of them is fake, I think it will be proven that this section is fake due to the highest correlation. What are these three hig h-level monitoring?

Detailed Predictions #

1. AI becomes a ubiquitous everywhere, just as online is ubiquitous.
2. We have not created an artificial spiritual system, and progress in any field is still dependent on the following: Overview of the goal, the possibility of using the time axis, data required for thinking (this)). In order).
3. It is difficult to measure the degree of terminal theory, but hopefully the first sentence does not give too much room for interpretation. For example, if it is known that the US military will use a drone controlled by artificial intelligence without human participation in the assassination decision by 2050, this will definitely be Jesus. 。
4. Language mode l-based chatbots are widely used in these applications as technical support and assistants like Siri and Hey Google. These assistants are ready to execute these tasks, such as reservations for airline tickets and requests for products on Amazon, but problems with deviations (the most difficult 5-10 %) in these tasks. Will occur. In other sentences, the more people the human assistant is, the more scary they are.

Biology (research) #

1. Robot engineering has finally begun to provide the benefits of bots improving the productivity of the manufacturing industry. Last but not least, the bot is still not ready to do an important level of "dirty" work to be widely accepted. For example, 95 % of the raving of the hedge is still being done.
2. A system that originated in artificial intelligence (assumed to be) accepts military conclusions in the United States with a limited or absolute or absolute existence in the circuit. This is not an endory thing, as seen in a video clip on massacre, but it leads to a mistake (looking back), and the average focus of human sensual separation from war and incidents. Get it worse.
3. In synergy with VR, we will have our own individual avatar based on images and tracking of our movements, set to support the origin of artificial minds 100% rather than a more lasting appearance than we have today. Overall, art will become even more democratic, and with the support of generative models, literally anyone will be able to create works of art of professional significance, which are linked together with images by video.

Biotech #

1. Progress in the field of GWAS continues. Difficulties regarding the non-European origin of presenters have been resolved, at least for other large (non-European) populations. The prognostic power of polygenic risk traits continues to grow, indeed for difficult diseases, thanks to the emergence of EQTL strategies and the final combination of the first biobank, which connects more than twice as many people as the UKBB.
2. ML models are being widely distributed as tools in the kit of biotools, which are presented in this excellent blog post.
3. Protein engineering is becoming much easier thanks to the combination of highly standardized work processes, more perfect tools with ML, and the costs of synthesis and (possibly including protein) sequencing.~Universe
4. Universe

Healthcare & healthtech #

1. Universe
2. 2%) in real time, and a large share in disease areas unrelated to cancer. (The awareness of these treatments is even higher: the information is there, but the experts don't understand, for example, what it is now (chemotherapy).
3. Gene therapy has guaranteed (approved) treatments, including 10 or even 100 genetic diseases that cannot be treated in real time. Some gene therapies are approved to cure unpopular diseases, such as cancer, heart disease, Alzheimer's, Parkinson's, etc.
4. OMIC screens combined with biomarker models are becoming more widespread, though not widespread. Read the article "Current Dive West" to find out what it is and what the problems are. In the US, this will make the healthiest people even healthier and more conscious of their health, while those who do not actively try to improve their health will not benefit at all.
5. Artificial intelligence systems will be used routinely by pathologists and radiologists, but they will not replace them completely. That is, in my opinion, the number of radiologists will not be reduced to more than half what it is today.~{space}
6. {space}
7. Space~The obesity epidemic has either peaked or is starting to go nowhere, and if it is the latter, it has somewhat. This is mainly due to the continued development of new generations of weight-reducing drugs. It is likely that obesity will fall below 40% again, but will remain above, say, 30%.
8. The fight against cancer is not "won" in the sense that more than a million people will still die from cancer in the United States. Meanwhile, the incidence of cancer continues to fall, with deaths at least halved compared to today. Unfortunately, cardiovascular disease is also still the leading cause of death in America. (In some ways, both projections are quite optimistic.
9. Space
10. {Ludge}
11. {Ludge}
12. overview

Synthetic Biology #

1. range
2. {Ludge}
3. rumor
4. {RUCE}~Most of the complex creatures de novo design and engineering (contrasting De Novo synthesis, which have modified existing genome) remain science fiction. Companies are still focusing on biological production of generic materials and compounds and engineering of mammal cells for treatment.

Crypto #

1. There are several for bios production (
2. {space}
3. {space}

Things I decided not to predict about #

The market price of the main blockchain / cryptocurrency token is higher than the main metal market price (2050).

1. At the same time, one nation got a digital currency unit from the central bank.
2. universe

The impact of AI on law, advice and money. Initially, I wanted AI to affect this field, but for example, it would not have a solid effect while you were waiting simple. However, it turned out that it was unimaginable to actually operate this.