Vitalik’s latest article: After a year, review D/ACC

Author: Vitalik, founder of Ethereum; compilation: 0xjs@作作作作作作作

About a year ago, I wrote>An article about technical optimismDescribes my enthusiasm for technology and the great benefits it can bring, and my caution of some specific issues. These concerns are mainly concentrated on Super Smart AI, and if the technology is built in a wrong way, it may bring it brings to the wrong way.Doom or irreversible human beings are powerless.One of the core ideas of my article is: decentralization, democratization, and differentiation defense acceleration (Decentralized and democratic, differential defensive accelerationTo.Accelerate the development of technology, but we must distinguish it, focus on the technology that improves our defense ability, rather than improving our ability to cause damage, and the technology of decentralized power, rather than focusing on power on behalf of everyone to decide, false and evil, good and evil,Single elite’s hands.Defense is like Democratic Switzerland and historically quasi -governmental Zhemia (Note: James C. Scott, a famous American sociologist, study the concept of Zomia proposed by Southeast Asia.Feudal lords and castles.

In the year after that, my ideas and ideas have matured a lot.I am here>>>>>>>>>>>>>>>>>>>>>>>>>>80,000 HoursWhen talking about these ideas, I saw a lot of response, most of them were positive and criticized.This work itself is continuing and achieved results: we see that it can be verifiedOpen -source vaccine progress, People are more and more recognize the value of indoor air, and Community Notes continues to shine,>The predicted market as an information toolThe year of breakthrough,>Government IDandSocial mediaZK-SNARK (and byAccount abstraction ProtectEthereum),,Open source imaging toolApplications in medicine and BCI, and so on.In autumn, we organized the first important D/ACC activity:Discovery Day (D/Addy) in DEVCON’s “D/ACC)There are speakers from all pillars (biology, physics, network, information defense, and neural technology) from D/ACC all day.People who have been committed to these technologies have been more and more aware of each other’s work, and outsiders have become more and more aware of greater stories: stimulating the same values of Ethereum and cryptocurrencies can be applied to a wider world.

Directory of this article

1. What is D/ACC, not what
Second, third dimension: survival and development
Third, problems: AI security, urgent timeline and supervision
Fourth, the role of cryptocurrencies in D/ACC
5. D/ACC and public goods financing
6. Future Outlook

1. What is D/ACC, not what

It is 2042.You see reports about new epidemic diseases in your city in the media.You have become accustomed to this situation: people are excited about the mutation of each animal disease, but most of them are not available.The previous two actual potential epidemic diseases were discovered very early through wastewater monitoring and open source analysis of social media, and they completely prevented their spread.But this time, the forecast market shows that the possibility of at least 10,000 cases is 60%, so you are more worried.

The sequence of the virus was determined yesterday.The software update of the portable air detector is already available, and new viruses can be detected (one breathing, or exposed in indoor air for 15 minutes).The open source description and code of the equipment to produce vaccines that can be found using any modern medical institution in the world should be available within a few weeks.Most people have not taken any actions, mainly relying on air filtration and ventilation to protect themselves.You have immune diseases, so you are more cautious: your open source local personal assistant AI, it handles navigation, restaurant and activity recommendation tasks. It also considers real -time air tester and carbon dioxide data. It only recommends the safest place.Data are provided by thousands of participants and devices, using ZK-SNARK and Differential Privacy to maximize the risks of data leakage or abuse of any other purposes (if you want to contribute data to these data sets, there is stillOther personal assistants AI can verify that these encrypted small tools are indeed effective formal proof).

Two months later, the epidemic disappeared: 60% of people followed the basic procedures, that is, if the air detector sent a beep and showed a virus, we put on a mask. If their personal testing is positive, they will stay.At home, this is enough to push the spread rate (because of the passive severe air filtering, which has been greatly reduced) to less than 1.Simulation shows that this disease may be five times more serious than COVID 20 years ago, but it is not a problem today.

DEVCON D/ACC Day

One of the most active gains on the D/ACC activity on DevCon is the successful cluster of D/ACC’s umbrella structure to successfully gather people from different fields and make them really interested in each other’s work.

It is easy to create a “diversity” activity, but it is difficult to make people with different backgrounds and interests really connect with each other.I still remember being forced to watch long opera in middle school and high school, and I personally think they are boring.I know I “should” appreciate them, because if I don’t appreciate it, then I will become a lazy computer science without culture, but I don’t have to connect with content at a more real level.D/ACC Day does not have this feeling at all: I feel that people actually like to learn different types of work in different fields.

If we want to create a more bright alternative than rule, deceleration and destruction, we need this extensive alliance construction.The D/ACC seems to be successful, and it is enough to explain the value of this idea.

D/ACC’s core ideas are simple: Decentralized and Democratic Differential Defensive Acceleration (Decentralized and Democratic Difference).Establish a technology to turn offensive/defensive balance to defense, and does not rely on transferring more power to the central authority.There is an internal connection between the two: any decentralization, democracy or freedom of political structure will flourish when the defense is easy, and the biggest challenge will be suffered when defense difficulties -in these circumstances, there are even more situations, even more, there are even more.The possible result is that everyone opposes everyone’s war period, and ultimately realizes the balance of strong rule.

The core principles of D/ACC cover many areas:

The chart comes from my article last year “>My technical optimism “”

One way to understand the importance of working hard to achieve decentralization, defensiveness and acceleration is to compare it with the philosophy obtained by abandoning the three.

Decentralization acceleration, but do not care about the “differentiated defense” section.Basically, becoming an E/ACC, but decentralization.Many people adopt this method. Some people mark themselves as D/ACC, but they describe their focus as “offensive”, but many people are excited about “decentralized AI” and similar themes in a more gentle way.But in my opinion, they do not pay enough attention to “defense”.
In my opinion, this method may avoid the global dictatorship of a specific tribe that you worry about, but it cannot solve the potential structural problem: there is a continuous disaster risk in an environment that is conducive to offense, or someone will beIt is positioned as a protector and permanently established his position.In the specific circumstances of AI, it cannot solve the risk of being deprived of power compared with AI as a whole.
Differential defensive acceleration, but does not care about “decentralization and democratization.”Embracing centralized control for a small number of people has a permanent attraction for a small number of people. Readers are undoubtedly familiar with many examples and their shortcomings.RecentThe continuation of the equipment, such as the ankle label used as a substitute for prison in several countries … encrypted videos and audio are continuously uploaded and explained by the machine in real time. “However, centralized control is a scope.A centralized control version that is usually ignored but still harmful is a resistance to public supervision in the field of biotechnology (such as food, vaccine), and allowing such a closed source of non -challenging.
Of course, the risk of this method is that centralization itself is often the source of risks.We have seen this in the new crown epidemic. The functional acquisition of the function funded by the major government of the world may be the source of the epidemic. The centralized understanding has led the WHO for many years.The requirement of mandatory social isolation and vaccination requires political rebounds that may last decades.Any risk related to AI or other high -risk technologies may occur.The method of decentralization can better cope with the risk from the centralization itself.
Decentralization defense, but do not care about acceleration—— Basically, try to slow down technological progress or economic recession.
The challenge of this strategy is dual.First of all, in general, technology and economic growth are of great benefit to humans, and any delay will>bringdifficult>EstimateThe price.Secondly, in a world of non -activeism, it is unstable: who has the most “cheating” and finds a seemingly reasonable way of progress and who will lead.The deceleration strategy can play a role to a certain extent in some cases: European food is healthier than American food, it is an example, and the success of non -diffusion so far is another example.But they cannot always be effective.

Through D/ACC, we hope:

In the era when most parts of the world becomes blunt, we must adhere to principles instead of building casually -On the contrary, we need to build some specific things to make the world safer and better.
Acknowledging index -level technological progress means that the world will become very strange, and the “footprint” of human beings in the universe will continue to increase.We must improve the ability to protect fragile animals, plants and humans from harm, but the only way out is to move forward.
Development technology that can ensure our security, rather than assume that “good people (or good AI) are in power”.We can implement this by developing tools. These tools are used for construction and protection>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>natureIt is more effective than when used to destroy.

Another way to think about D/ACC is to review a framework of the European Pirate Party Movement in the late 2000: empowerment.

Our goal is to build a world that retains human autonomy, which must achieve negative freedom and avoid active intervention (whether from other people from other people, from the government, or super intelligent robot).It is necessary to achieve positive and freedom to ensure that we have knowledge and resources.This echoed the classical liberal tradition of several centuries, including Stewart Brand’s attention to “acquisition tools”.Emphasis on the key component of progress — maybe, people can add a little, BuckMinster Fuller hopes to see that the global solution is involved and widely distributed.Considering the technical pattern of the 21st century, we can regard D/ACC as a way to achieve these same goals.

Second, third dimension: survival and development

In my article last year, D/ACC pays special attention to defense technology: physical defense, biological defense, network defense and information defense.However, decentralization defense is not enough to make the world better: you also need a forward -looking positive vision to see what humans can use its new discovery and security to achieve.

Last year’s article did contain positive views, manifested in two places:

In response to the challenges faced by super intelligence, I proposed a way that can have super intelligence without deprivation of power (this is far from my original):

Today, AI is made as a tool, not a highly autonomous smart body
Tomorrow will use virtual reality, muscle electrical signals and brain interface tools to establish increasingly close feedback between artificial intelligence and humans
Over time, the end result is that super intelligence becomes a combination of machines and our close combination.

When talking about information defense, I also mentioned by the way that in addition to trying to help the community maintain cohesion in front of the attacker and carry out high -quality discussion, there are some progressive social technology to help the community can make it easier for the communityHigh -quality judgment:pol.isIt is an example that predict the market as another example.

But these two points seem to be disconnected from the D/ACC’s argument: “There are some ideas that create a more democratic and defensive world at the basic layer. By the way, there are some irrelevant related to how we realize super intelligence.idea”.

However, in fact, there are some very important connections between the “defensive” and “progressive” D/ACC technology mentioned above.Let’s expand the D/ACC chart in the article last year, add this axis to the chart (similar, let us re -marked it as “survival and prosperity”), and see what the results will appear:

In all fields, there is a consistent model that can help us “survive” science, thoughts and tools in one field closely related to the science, thought and tools that can help us “prosper”.Here are some examples:

Recently, many research on anti -new coronal virus pays attention to the durability of the virus in the body, which is a reason why the new crown is so serious.Recently, there are also signs that the durability of the virus may be the cause of Alzheimer’s disease -if it is true, it means that the persistence of solving the virus in all tissue types may be the key to solving aging problems.
Some of the low -cost micro imaging tools developed by OpenWater can effectively treat micro -thrombosis, continuous existence of viruses, and cancer, and can also be used for BCI.
Very similar ideas have prompted people to build social tools that are highly constructed by environmental construction, such as community notes, and social tools built for a reasonable cooperation environment, such as POL.IS.
The prediction market is very valuable in high -level cooperation and high adversity environment.
Zero -knowledge proof and similar technology that calculates data while protecting privacy increases data that can be used for useful work such as science, but also enhance privacy.
Solar and batteries are of great benefit to promoting the next wave of clean economic growth, but they also perform well in decentralization and physical flexibility.

In addition, there are also important cross -dependence relationships between the fields of disciplines:

BCI is very important as an information defense and collaboration technology because it allows us to exchange our ideas and intentions in detail.BCI is not only a communication between robots and consciousness, it can also be a communication between consciousness and robotic and consciousness.This and>Pluality’s concept of BCI valueEcho.
Many biotechnology depends on information sharing. In many cases, people are willing to share information only when they are convinced that information will be used for an application and only for an application.This depends on privacy technology (such as ZKP, FHE, OBFUSCATION …)
Collaborative technology can be used to coordinate funds in any other technical fields

Third, problems: AI security, urgent timeline and supervision

Different people have a completely different AI timetable.The chart comes from ZUZALU in Heishan, 2023.

I found that the most convincing opposition to my article last year was criticism from the AI security community.This argument is: “Of course, if we can get a strong AI for half a century, we can focus on creating all these good things. But in fact, we are likely to have three years to achieve AGI.Only when we can realize super intelligence, so if we do not want the world to be destroyed or caught into an irreversible trap, we cannot just accelerate the occurrence of good things, and we must slow down the occurrence of bad things.It will feel uneasy about people with power. “In my article last year, I did not call for any specific strategies to” slow down the occurrence of bad things “, but it was vaguely called not to build a risky super intelligent form.Therefore, here is worth answering this question directly: If we live in the most inconvenient world, the risk of artificial intelligence is high, and the timetable may take five years. What regulations will I support?

First of all, be cautious about new regulations

Last year, the main artificial intelligence regulatory bill proposed by California was the SB-1047 Act.SB-1047 developers with the most powerful models (those models spent more than 100 million US dollars for training or fine-tuning of more than 10 million US dollars) have taken some security test measures before their release.In addition, it also stipulates that if developers of artificial intelligence models are not cautious enough, they have to take responsibility.Many critics believe that the bill is “threat to open source”; I disagree because the cost threshold means that it only affects the most powerful model: even LLAMA3 may be lower than the threshold.However, when I look back, I think there is a bigger problem in the bill: like most regulations, it is overly matched with today’s situation.In the face of new technologies, the attention of training costs has become fragile: the training cost of the most advanced DeepSeek V3 model recently is only $ 6 million, but in the portraitO1In such a new model, costs are shifting from training to more common reasoning.

Secondly, participants who are most likely to truly be responsible for artificial intelligence super intelligent destruction scenes are actually military.As we have seen in the field of biological safety in the past half century (and later), the army is willing to do terrible things, and they are easy to make mistakes.Artificial intelligence military applications are rapidly developing (see Ukraine, Gasha).Any security regulations adopted by the government default to exempted from their national army and companies that work closely with the military.

Nevertheless, these arguments are not the reason we give up and do not take any action.On the contrary, we can use it as a guidance and try to formulate the rules that cause at least these concerns.

Strategy 1: Responsibility

If someone’s behavior causes damage to the law, they may be prosecuted.This does not solve the risk problem from the army and other actors from the army and other actors, but it is a very common method that can avoid excessive fitting, and it is usually suffered from liberalism for economics for this reason.Home support.

The main goal of the responsibility to consider so far is:

Users -people who use artificial intelligence
Deployee -intermediaries that provide users with AI services
Developers -people who build artificial intelligence

It seems that it seems that the responsibility is the most consistent with the incentive mechanism.Although the connection between the model development method and the final usage method is usually unclear, users determine the specific use of artificial intelligence.Letting users take responsibility will put strong pressure on artificial intelligence, forcing people to develop artificial intelligence in the correct way I think: focusing on creating a mech suit for human thinking, instead of creating a new form of self -maintaining intelligence.The former will regularly respond to the user’s intention, so it will not cause catastrophic behaviors unless the user wants to do so.The latter is most likely to get out of control and cause a typical “artificial intelligence out of control” scene.Another advantage of liability as close to the final location is that it can minimize responsibilities and cause people to adopt the risk of other harmful behaviors (for example, closed sources, KYC and monitoring, national/enterprise collusion to restrict users, such as canceling cancellationBank business, blocking most of the world).

There is a classic argument that blames the user completely on the user: users may be ordinary people who do not have much money, even anonymous, and no one can really pay for catastrophic damage.This argument may be exaggerated: even if some users are too small and do not need to bear responsibility, ordinary customers of artificial intelligence developers are not like this, so artificial intelligence developers will still be inspired to buildProducts of responsibility risks.Having said that, this is still an effective argument that needs to be solved.You need to motivate people who have resources in the pipeline to take appropriate cautious measures, and deployers and developers are easy to be attacked. They still have a great impact on the security of the model.

The responsibility of the deployment seems to be reasonable.A common concern is that it is not suitable for open source models, but this seems to be controlled, especially because the most powerful model is likely to be closed (if they are eventually open source, although the deployee’s responsibility is ultimateIt is not very useful, but it will not cause much harm).Developers have the same concern (although the open source model needs to fine -tun the model to make it performing some things that are not allowed at the beginning), but the same refutation is also applicable.As a general principle, the collection of “tax” for control is essentially “you can build things you can’t control, or you can build things you control, but if you build things you control, then 20% of the control must be used”In our purpose”, this seems to be a reasonable position of the legal system.

An idea that seems to be fully explored is to blame other participants in the pipeline, and they can better guarantee sufficient resources.A very conducive to D/ACC’s idea is that in the process of performing certain catastrophic harmful actions, blame the owner or operator of any device who takes responsibility for artificial intelligence to take over (such as a hacker attack).This will produce a very wide incentive, prompting people to work hard, so that the infrastructure of the world (especially computing and biological) is as safe as possible.

Strategy 2: The global “soft pause” button on industrial -grade hardware

If I am convinced that we need to be more “powerful” than the rules of responsibility, this is what I want to pursue.The goal is to reduce the global useful computing power in a critical period to approximately 90-99% to 1-2 years, and strive for more time for human beings to prepare.The value of 1-2 years should not be exaggerated: in the case of complacent, the one-year “wartime mode” is easy to get off the work of one hundred years.It has explored the implementation of “suspension”, including specific suggestions such as registration and verification of hardware positions.

A more advanced method is to use clever encryption techniques: for example, the industrial grade (but not consumer) artificial intelligence hardware can be equipped with a credible hardware chip. Only a week from major international institutions (including at least one non -military affiliated institution) Get 3/3 signatures to continue running.These signatures will have nothing to do with the device (if necessary, we can even ask zero -knowledge proof to prove that they have been published on the blockchain), so this will be all or not: there is no actual method.The device is authorized to continue running a device.

This feels like “checking the box” in terms of maximum interests and minimized risks:

This is a very useful ability: if we receive a warning signal, it shows that the almost super intelligent AI starts to do something that may cause disaster destruction, and we will want to change slowly.
Before this critical moment, it will not cause much harm to developers alone.
Focusing on industrial scale hardware, and only 90-99%, can avoid the practice of mounting spy chips or terminating switches in consumer laptops, or forcing small countries to take severe measures such as strict measures.
Focusing on hardware seems to be very effective for technical changes.We have seen that in the multi -generation artificial intelligence, the quality depends to a large extent on its available computing power, especially in the early version of the new paradigm.Therefore, the useful computing power will be reduced by 10-100 times, and super-intelligent artificial intelligence, which is easy to decide out of control, won or failed in a fast-paced battle with humans trying to stop it.
It is annoying to sign a signature every week itself, which will put a strong pressure on the expansion of the plan to the consumer hardware.
It can be verified by random inspections, and verification at the hardware level will make it difficult for specific users to exempt (based on law compulsory shutdown instead of technical methods without such full or completely unsatisfactory attributes, which makes them even moreIt is possible to slide to the army and other exemptions)

Hardware supervision has been seriously considered. Although it is usually a framework for export control, export control is essentially a concept of “we trust us, but not trust the other party”.Leopold Aschenbrenner has advocated that the United States should compete for a decisive advantage, and then forced China to sign an agreement to limit the number of boxes they allowed to run.In my opinion, this method seems to be dangerous, and it may combine the defects of multi -pole competition and centralization.If we have to limit people, it seems that it is best to limit everyone on the basis of equality, and try to try to organize this job, rather than trying to dominate other people.

D/ACC technology in AI risk

These two strategies (responsibilities and hardware pause buttons) have loopholes, and it is clear that they are just temporary right measures: if something can be completed on the super computer at T time, then after T + 5 years, it is very very veryIt may be completed on the laptop.Therefore, we need more stable things to fight for time.Many D/ACC technologies are related to this.We can look at the role of D/ACC technology like this: What will it do if artificial intelligence takes over the world?

It invades our computer → network defense
It can cause super plague → biological defense
It allows us to convince us (either believe in it or not trust each other) → information defense

As mentioned above, the rules of responsibility are a natural D/ACC -friendly supervision method, because they can effectively inspire the use of these defensive measures around the world and take them seriously.Taiwan has recently been experimenting with the responsibility of false advertising, which can be regarded as an example of the use of responsibility to encourage information defense.We should not be too passionate to apply responsibilities anywhere, and remember the benefits of ordinary freedom, that is, let the small characters participate in innovation without having to worry about litigation, but in the place where we really want to promote security more strongly, the responsibility can be very flexible and effectiveEssence

Fourth, the role of cryptocurrencies in D/ACC

Most of D/ACC’s content far exceeds the typical blockchain theme: biological safety, BCI and collaborative discourse tools seem to be far from cryptocurrency people.However, I think there are some important connections between cryptocurrencies and D/ACC, especially: especially:

D/ACC is an extension from the underlying value of encrypted (decentralized, anti -review, open global economy and society) to other technical fields.
Because encrypted users are naturally adopted and the values are consistent, the encrypted community is a natural early user of D/ACC technology.Pay attention to the community (including online and offline, such as activities and pop -up windows), and these communities are actually doing high -risk things rather than talking about each other, making the encrypted community a particularly attractive D/ACC technology incubators and test tables, these technologies are fundamentally targeted at groups rather than individuals (for example, a large part of information defense and biological defense).The encrypted people are just doing things together.
Many encrypted technologies can be applied to D/ACC theme fields: blockchains used to build more powerful and decentralized financial, governance and social media infrastructure for zero -knowledge certificate for privacy.Today, many of the largest forecast markets are based on the blockchain, and they are gradually becoming more complicated, decentralized and democratic.
There are also win -win opportunities to cooperate in terms of encryption -related technologies. These technologies are very useful for encrypted projects, but they are also the key to achieving D/ACC targets: formal verification, computer software and hardware security and strong confrontation technology.These things make the Ethereum blockchain, wallet, and DAO more secure and powerful. They also achieve important civilized defense goals, such as reducing our fragility against network attacks, including possible attacks from super smart AI.

CURSIVE is an application that uses a full -state encryption (FHE) that allows users to identify areas that are interested in with other users while protecting privacy.Zuzalu, one of Chiang Mai’s many branches, uses this technology.

In addition to these direct intersections, there are another important common interest point: capital mechanism.

5. D/ACC and public goods financing

One of my always interesting mechanisms to think of better mechanisms to fund public products: items that are valuable to a large number of people but do not have a natural and feasible business model.My past work in this area includes my contribution to secondary financing and its application in Gitcoin, Retro PGF, and recent deep financing applications.

Many people are skeptical of the concept of public items.This suspicion usually comes from two aspects:

In fact, public goods have been used in history as a strict central plan and the government’s excuse for intervention in society and economy.
It is generally believed that the lack of rigor for public goods is based on social expectations (it sounds good, not actually good), and favors insiders who can play social games.

These are important criticism and good criticism.However, I think that strong decentralized public goods financing is very important for D/ACC’s vision, because a key goal of D/ACC (minimizing the central control point) will essentially frustrate many traditional business models.It is possible to build a successful company on the basis of open source -many Balvi recipients are doing this -but in some cases, it is difficult enough to be difficult, so that important projects require additional continuous support.Therefore, we must do difficult things and find out how to solve public goods financing in a way to solve the above two criticisms.

The solution of the first problem is basically credible neutral and decentralized.There is a problem with the central plan, because it gives the control to elites that may abuse their powers, and it often coincides with the current situation. Over time, its efficiency is getting lower and lower.The secondary financing and similar mechanisms provide funds for public goods based on decentralized (architecture and politics) decentralized (architecture and politics) as possible.

The second problem is more challenging.For secondary financing, a common criticism is that it will soon become a popular competition, asking project funders to spend a lot of energy to publicize publicity.In addition, those “in front of people” (such as end -user applications) will receive funding, but those more hidden projects (typical “dependent dependence by people in Nebraska”)Fund.Optimism’s RETRO financing relies on a few expert badges holders; here, the impact of popularity competitions has been weakened, but the social impact of maintaining close personal relations with the owner of the badge is amplified.

Deep financing is the latest effort I made to solve this problem.There are two main innovations in deep financing:

Dependency diagram.We will not ask each juror as a global question (“What is the value of the project A to humans?”) Instead, it raises a local question (“Is project A or Project B more valuable to results C? How much is the value high?”)).As we all know, humans are not good at answering overall questions: In a famous study, when asked how much they are willing to spend to save N, the interviewees answered N = 2,000, N = 20,000, and N = 200,000 answers.About $ 80.Local problems are easier to deal with.Then, we combine local answers into a global answer by maintaining the “dependency relationship”: What other projects have contributed to each project and what is the degree of contribution?
Artificial intelligence is a distilled human judgment.The jurors will only be assigned to a small part of all issues of all issues.There is a public competition that can submit an artificial intelligence model to try to effectively fill all the sides in the map.The final answer is the weighted summary of the most compatible model with the jury.SeeEssenceThis method allows the mechanism to expand to a very large scale, and at the same time, the jury is required to submit only a small amount of information “bit”.This reduces the chance of corruption and ensures that each one is high -quality: jurors can think of each problem for a long time, rather than quickly click on hundreds of questions.By using artificial intelligence public competition, we have reduced any bias of any single artificial intelligence training and management process.The open market of artificial intelligence is engine, and humans are steering wheels.

But in -depth funding is only the latest example; there have been the idea of other public goods funding mechanisms in the past, and there will be more in the future.Allo.expertrightThey have a good classification.Its fundamental goal is to create a small social tool that can fund public items, and its accuracy, fairness and openness are at least close to the way of funding private items.It does not have to be perfect; after all, the market itself is far from perfect.However, it should be effective enough to work hard to benefit everyone’s top -level open source projects to continue to do this without having to feel that it is necessary to make unacceptable compromises.

Nowadays, most D/ACC theme areas are the leading projects: “Edge BCI”, anti -aging drugs, hardware, etc. are proper.This has great shortcomings in ensuring public trust, as we have seen in many areas.It also shifted its attention to the competitive developments (“Our team must win this key industry!”) Instead of greater competition, that is, to ensure that these technologies have developed fast enough to protect us in the super smart AI world.For these reasons, strong public goods funds can strongly promote openness and freedom.This is another way for the encrypted community to help D/ACC: by earnestly explore these financing mechanisms and enable them to play a good role in their own environment, and prepare for more extensive adoption of open source science and technology.

6. Future Outlook

In the next decades, major challenges will be faced.I have been thinking about two challenges recently:

The strong wave of new technologies, especially strong artificial intelligence, is emerging rapidly, and these technologies are accompanied by important traps we need to avoid.”Artificial Intelligence Super Smart” may take five years to achieve, or 50 years.In any case, it is unclear whether the default result must be positive. As described in this article and the previous article, there are multiple traps to be avoided.
The world is increasingly lacking cooperation.Many powerful behaviors seem to be at least at least in the past (worldism, freedom, common humanity, etc.), but now they are more open and actively pursuing the personal interests of individuals or tribes.

However, these challenges have a first -line hope.First of all, we now have very powerful tools to complete the remaining work faster:

At present and not far, AI can be used to build other technologies and can be used as a component of governance (such as deep financing or>Information financeTo.It is also very related to BCI, and BCI itself can further improve productivity.
Large -scale coordination is now more likely than before.The Internet and social media have expanded the scope of coordination. Global finance (including cryptocurrencies) has enhanced its strength. Now information defense and collaborative tools can improve their quality. Perhaps soon human -machine BCI can increase its depth.
Formation verification, sandbox (network browser,DockerAs well asQubesAs well asGrapheneosWait), security hardware modules and other technologies are constantly improving, so that better network security can be achieved.
Writing any type of software is much easier than two years ago.
Recently, the basic research on the mechanism of virus action, especially the most important form of prevention and control of virus transmission is the simple understanding of air spread, and pointed out a clearer path for how to improve biological defense capabilities.
The latest progress of biotechnology (such as CRISPR, progress in biological imaging) is making various biotechnology easier to achieve, whether it is for defense, longevity, super happiness, or exploring a variety of novel biological assumptions, or just doing itSome cool things.
Computing and the advancement of biotechnology have led to the emergence of synthetic biological tools. You can use these tools to adjust, monitor and improve your health.Code science and other network defense technology make personalization more feasible.

Secondly, since the many principles we cherish are no longer occupied by a few conservatives, they can be regained by a widely -known alliance that welcomes anyone in the world.This may be the biggest benefit of the recent “re -adjustment” of politics around the world and is worth using.Cryptocurrencies have already used this and gained global attraction; D/ACC can also do this.

Obtaining tools means that we can adapt to and improve our biology and environment, and the “defense” part of D/ACC means that we can do this without infringing the freedom of others.>Liberty and Diversity PrinciplesIt means that we can have a lot of diversity in this regard, and our commitment to common human goals means that it should be realized.

We humans are still the most dazzling stars.The task we face is to build a more bright 21st century and protect human survival, freedom and autonomy in the process of moving towards the starry sky. This is a challenging task.But I believe we can do it.