The Bitter Lesson by Rich Sutton

In his influential essay, "The Bitter Lesson", Rich Sutton, a prominent figure in reinforcement learning, argues that the most significant insight from 70 years of AI research is the ultimate triumph of general-purpose methods that leverage computation over those that rely on incorporating human knowledge. Sutton posits that while building in domain-specific human knowledge can provide short-term gains, these approaches tend to plateau and even impede long-term progress. In contrast, methods that scale with increasing computational power, such as search and learning, have consistently led to breakthroughs.

Sutton supports his argument with several key examples from the history of AI:

• Computer Chess: Early attempts to create chess-playing programs focused on encoding human strategies and knowledge. However, the system that ultimately defeated world champion Garry Kasparov in 1997, Deep Blue, was based on massive, deep search capabilities.

• Computer Go: Similarly, in the game of Go, initial efforts to leverage human understanding of the game were surpassed by systems like AlphaGo, which relied on search and learning from self-play.

• Speech Recognition: The field of speech recognition saw a shift from knowledge-based systems to statistical methods like Hidden Markov Models (HMMs), which performed significantly better in a 1970s DARPA competition. The more recent success of deep learning in this area further underscores the power of computation and learning from large datasets.

Sutton's "bitter lesson" is a four-part observation: 1) researchers build knowledge into their agents, 2) this provides a short-term boost, 3) it ultimately plateaus and hinders further progress, and 4) breakthroughs consistently come from scaling computation with search and learning. He concludes by advocating for the development of meta-methods that can discover and capture the complexity of the world on their own, rather than being explicitly programmed with human discoveries.

The Bitter Lesson's Bitter Lesson by Andrew Trask

"The Bitter Lesson's Bitter Lesson" presents a critique and extension of Sutton's argument. Trask contends that Sutton's focus on "pure learning" from scratch, akin to how babies and animals learn, is computationally impractical and overlooks the immense value of "inherited learning" from human-generated data.

Trask introduces several key quantified points to support his argument:

• The Scale of Evolution: Trask estimates that the evolutionary process that produced human intelligence involved over 10^50 operations. In contrast, current state-of-the-art AI models are trained with around 10^26 operations. This vast difference suggests that recreating the learning process from scratch is computationally infeasible.

• The Efficiency of Inherited Learning: Trask argues that human-generated text is a highly compressed and efficient source of knowledge, representing the output of 4.5 billion years of evolutionary optimization. By learning from this data, AI models can inherit a massive amount of information without having to rediscover it.

• Untapped Human Data: While some may believe that large language models (LLMs) have consumed the entire internet, Trask points out that the training datasets of leading AI models are in the range of 100-200 terabytes. However, the total amount of digitized human data is estimated to be around 180 zettabytes. This means that current AI models are using less than a millionth of the available human-generated data.

Trask's central thesis is that the future of AI lies in developing architectures that can effectively and privately access this vast, untapped repository of human knowledge. He argues for a hybrid approach that combines the benefits of inherited knowledge with the ability for novel discovery, moving beyond the limitations of "pure learning."

1. Do not obey in advance. Most of the power of authoritarianism is freely given. In times like these, individuals think ahead about what a more repressive government will want, and then offer themselves without being asked. A citizen who adapts in this way is teaching power what it can do.

2. Defend institutions. It is institutions that help us to preserve decency. They need our help as well. Do not speak of "our institutions" unless you make them yours by acting on their behalf. Institutions do not protect themselves. They fall one after the other unless each is defended from the beginning. So choose an institution you care about -- a court, a newspaper, a law, a labor union -- and take its side.

3. Beware the one-party state. The parties that remade states and suppressed rivals were not omnipotent from the start. They exploited a historic moment to make political life impossible for their opponents. So support the multiple-party system and defend the rules of democratic elections. Vote in local and state elections while you can. Consider running for office.

4. Take responsibility for the face of the world. The symbols of today enable the reality of tomorrow. Notice the swastikas and the other signs of hate. Do not look away, and do not get used to them. Remove them yourself and set an example for others to do so.

5. Remember professional ethics. When political leaders set a negative example, professional commitments to just practice become more important. It is hard to subvert a rule-of-law state without lawyers, or to hold show trials without judges. Authoritarians need obedient civil servants, and concentration camp directors seek businessmen interested in cheap labor.

6. Be wary of paramilitaries. When the men with guns who have always claimed to be against the system start wearing uniforms and marching with torches and pictures of a leader, the end is nigh. When the pro-leader paramilitary and the official police and military intermingle, the end has come.

7. Be reflective if you must be armed. If you carry a weapon in public service, may God bless you and keep you. But know that evils of the past involved policemen and soldiers finding themselves, one day, doing irregular things. Be ready to say no.

8. Stand out. Someone has to. It is easy to follow along. It can feel strange to do or say something different. But without that unease, there is no freedom. Remember Rosa Parks. The moment you set an example, the spell of the status quo is broken, and others will follow.

9. Be kind to our language. Avoid pronouncing the phrases everyone else does. Think up your own way of speaking, even if only to convey that thing you think everyone is saying. Make an effort to separate yourself from the internet. Read books.

10. Believe in truth. To abandon facts is to abandon freedom. If nothing is true, then no one can criticize power, because there is no basis upon which to do so. If nothing is true, then all is spectacle. The biggest wallet pays for the most blinding lights.

11. Investigate. Figure things out for yourself. Spend more time with long articles. Subsidize investigative journalism by subscribing to print media. Realize that some of what is on the internet is there to harm you. Learn about sites that investigate propaganda campaigns (some of which come from abroad). Take responsibility for what you communicate with others.

12. Make eye contact and small talk. This is not just polite. It is part of being a citizen and a responsible member of society. It is also a way to stay in touch with your surroundings, break down social barriers, and understand whom you should and should not trust. If we enter a culture of denunciation, you will want to know the psychological landscape of your daily life.

13. Practice corporeal politics. Power wants your body softening in your chair and your emotions dissipating on the screen. Get outside. Put your body in unfamiliar places with unfamiliar people. Make new friends and march with them.

14. Establish a private life. Nastier rulers will use what they know about you to push you around. Scrub your computer of malware on a regular basis. Remember that email is skywriting. Consider using alternative forms of the internet, or simply using it less. Have personal exchanges in person. For the same reason, resolve any legal trouble. Tyrants seek the hook on which to hang you. Try not to have hooks.

15. Contribute to good causes. Be active in organizations, political or not, that express your own view of life. Pick a charity or two and set up autopay. Then you will have made a free choice that supports civil society and helps others to do good.

16. Learn from peers in other countries. Keep up your friendships abroad, or make new friends in other countries. The present difficulties in the United States are an element of a larger trend. And no country is going to find a solution by itself. Make sure you and your family have passports.

17. Listen for dangerous words. Be alert to use of the words "extremism" and "terrorism." Be alive to the fatal notions of "emergency" and "exception." Be angry about the treacherous use of patriotic vocabulary.

18. Be calm when the unthinkable arrives. Modern tyranny is terror management. When the terrorist attack comes, remember that authoritarians exploit such events in order to consolidate power. The sudden disaster that requires the end of checks and balances, the dissolution of opposition parties, the suspension of freedom of expression, the right to a fair trial, and so on, is the oldest trick in the Hitlerian book. Do not fall for it.

19. Be a patriot. Set a good example of what America means for the generations to come. They will need it.

20. Be as courageous as you can. If none of us is prepared to die for freedom, then all of us will die under tyranny.