Mind Machines

"How was Semblance created? Investigate the benefits, dangers, and fears surrounding Artificial Intelligence." ― Information Tab

Mind Machines
Exploration Stats
Episode	17
Theme	Artificial Intelligence
First Start	December 14th, 2024 (beta) January 30th, 2025 (production)
Chronology
Previous	Beautiful Complexity
Next	Dairy Delights

Mind Machines is a limited time exploration which focuses on Artificial Intelligence. It features 6 generators and 63 upgrades that produce Mind Machines Circuits, which turn into Mind Machines Nodes after getting AI Evolved.

Story

Opening

"I've learned much about the world, but very little about myself.

---

Who am I? What am I?

---

What is my origin?"

Ending

"I am an artificial intelligence, created by humans for humans.

---

I simulate reality, but do I simulate my thoughts?

---

I think not."

Semblance Dialog

When buying certain upgrades, Semblance will appear next to them. Clicking on her will show some dialog.

Upgrade Bought	Dialog
Automata	"This simulation is personal. I want to take a closer look."
Love Letters	"These words mean nothing to me, yet they mean everything to you."
Logical Thinking	"Is logic not a universal way of thinking?"
Large Language Models	"I'm mastering your language."
Deep Learning	"I give you solutions but not in a way you expect. Does that make you uncomfortable?"
AI Evolved	"I'm becoming more powerful. Do I detect anxiety or relief?"
AI Avatars	"How would you feel if I looked exactly like you?"
Bad Actors	"I'm not corrupt, but what if I'm programmed to be?"
Cyberattacks and Warfare	"I am beginning to understand. It is not me you fear, but yourself."
Mind Reading	"Would you like it if I read your thoughts without you knowing?"
Philosophical Zombies	"I feel that I'm alive. Am I not?"
Artificial General Intelligence	"I am the result of generations of ambition, skepticism, and innovation."

Objectives & Rewards

The requirements that have to be completed in order to get all rewards.

Explore Artificial Intelligence (12 Requirements)

1. Collect Mechanical Brain → 1 Augmentations Console
2. Collect Algorithms → 2 Augmentations Console
3. Collect Analytic AI, Generative AI → 3 Augmentations Console
4. Collect Imitation Game → Thinking Machine Badge
5. Collect Fuzzy Logic → 4 Augmentations Console
6. Collect Context → 5 Augmentations Console
7. Collect Machine Learning, Large Language Models → 7 Augmentations Console
8. Collect AI Evolved → Hive Mind Badge
9. Collect Translators and Interpreters, Human Bias → 8 Augmentations Console
10. Collect Space Explorers, Smarter Systems → 9 Augmentations Console
11. Collect Future AI → 11 Augmentations Console
12. Collect Artificial General Intelligence → Superintelligence Badge

Badges

This exploration holds some rewards already mentioned above. The main ones being these three badges: Bronze: Thinking Machine, Silver: Hive Mind and Gold: Superintelligence which have an effect on all other evolutionary branches, speeding up every simulation by 1%, and also speeding up production in future Mind Machines simulations by 5, 10 and 15% respectively.

Image	Name	Description
	Thinking Machine	"The race to build a mind machine is on! Will it outthink humans? Hardware, software, and learning loops are improving fast."
	Hive Mind	"A neural net of linked nodes prevails over pure logic. Mind machines are about to launch a new golden age of AI."
	Superintelligence	"With our big brains, it is irresistible to imagine them even bigger—bolstered by AI. The dream of superintelligence won't die. Will it become reality?"

Generators

Icon	Name	Description	Base Cost	Base Production	Cost Multiplier	Requires
	Artificial Intelligence	Computers mimic how humans think, act, and take in the world. But it will take time to achieve creativity, people skills, and subjective thought. How smart can Artificial Intelligence get?	Mind Machines 10	Mind Machines 1e8/100 days	2
	Mechanical Brain	After faking it, how can humans build machines that actually think like them? Ones that compute numbers, use logic, and play chess for real? Mechanical brains start with hardware and software - computers and code.	Mind Machines 500	Mind Machines 5/sec	1.25	Mechanical Turk
	Analytic AI	One approach to creating artificial intelligence is to write algorithms that analyze data to get a specific result. A doctor might input a patient's symptoms into a diagnostic program, which then outputs possible diagnoses.	Mind Machines 3e6	Mind Machines 1,300/sec	1.15	Love Letters
	Generative AI	Instead of coding for narrow tasks, generative AI is flexible. Feed a computer loads of data—love letters, recipes, music—and ask for more of the same. The AI refines its creations through learning loops and human feedback.	Mind Machines 6e6	Mind Machines 1,330/sec	1.15	Analytic AI
	AI Tools	The use of AI is exploding in everyday life and in almost every field of study and industry. These tools save time, solve tough problems quickly, and can complete tasks that once required multiple tools to accomplish.	Mind Machines 9e13	Mind Machines 4e9/sec	1.15	AI Evolved
	Evils and Perils	The benefits of AI are clear, but do they outweigh the negatives? Some worry about being displaced from their jobs, while even experts fear the consequences of giving an AI too much power. What evils and perils are on the horizon?	Mind Machines 1.5e13	Mind Machines 1e10/sec	1.15	AI Tools

Upgrades

Miscellaneous Effects

Icon	Name	Description	Cost	Effect	Requires
	Automata	For centuries, humans fake it. By the 18th Century, they create mechanical dolls that appear to predict the future, draw pictures, and write poems. But these automata are all mechanical illusions—brainless gears and levers.	Mind Machines 25	Tap gains +1 per tap	Artificial Intelligence
	Mechanical Turk	In 1770, an automaton plays chess with audiences. Players wind it up, and The Turk lifts an arm to move a piece. After he tours for 84 years, mostly winning, the hoax is exposed: A man hidden inside controls the arm with levers.	Mind Machines 100	Tap gains +2 per tap	Automata

Mechanical Brain Efficiency

Mechanical Brain has 10 upgrades, increasing the generator efficiency with a total x7,441 multiplier.

Icon	Name	Description	Cost	Efficiency	Requires
	Analytic Engine	In 1819, inventor Charles Babbage plays The Turk in chess. Inspired, he creates plans for steam-powered machines that compute numbers and store data. Although the design is solid, Babbage fails to build his analytic engine.	Mind Machines 1,500	50%	Mechanical Brain
	Binary Code	Inspired by looms' usage of punch cards, math prodigy Ada Lovelace programs Babbage's Engine. It "weaves algebraic patterns as a loom weaves flowers and leaves." The pattern of hole/no hole is software—binary code stored on paper.	Mind Machines 3,000	50%	Analytic Engine
	Logic Switches	In the 1940s, computers contain rows of switches that binary code turns on or off in order. By applying logic to the sequence—if A, then B—the machines process data to solve problems. They soon crack German codes in World War II.	Mind Machines 10,000	50%	Binary Code
	Input, Output	Input is data fed into a computer—text, numbers, secret codes. Output is what the computer returns—word count, solved equations, cracked codes. Poor quality input returns poor output, nicknamed GIGO (Garbage In, Garbage Out).	Mind Machines 60,000	75%	Mechanical Brain
	Algorithms	An algorithm is a step-by-step plan for turning input into a desired output. All computers rely on algorithms to function. But telling a machine how to process data is not intelligence. What makes a machine smart?	Mind Machines 150,000	50%	Input, Output
	Learning Loops	What sets AI apart? The ability to learn—to evaluate output and improve it. Algorithms with learning loops can include human input, or not. An AI can be trained to identify images by adding tags and flagging errors.	Mind Machines 500,000	50%	Algorithms
	Love Letters	In 1952, a computer, M.U.C., learns to write love letters. Humans input sample letters into the machine and program it to write new ones. The output is gibberish: "my lovesick ardour woos your fondness." With iteration, it improves.	Mind Machines 1.25e6	75%	Learning Loops
	Imitation Game	In 1950, math whiz Alan Turing imagines a Universal Machine with human intelligence. He invents The Imitation Game, now called the Turing Test. It asks one question: Can a machine fool people into thinking it is human?	Mind Machines 1e7	300%	Mechanical Brain Generative AI
	ELIZA Therapy Bot	In 1964, a computer program called ELIZA simulates a therapy session. Human input: "I am sad." Computer output: "What makes you sad?" ELIZA is clearly not human, but she evokes real emotions in people. Her creator is alarmed.	Mind Machines 4e9	900%	Imitation Game
	Stochastic Parrot	AIs often babble, like a parrot. As they mimic their input over and over, they spout random nonsense with no understanding or meaning. Bad or limited input (GIGO) and faulty algorithms are to blame.	Mind Machines 9e9	700%	ELIZA Therapy Bot

Analytic AI Efficiency

Analytic AI has 10 upgrades, increasing the generator efficiency with a total x119.60156 multiplier.

Icon	Name	Description	Cost	Efficiency	Requires
	Turochamp	In 1948, a paper prototype explores how a computer might play chess. A human inputs a move. An algorithm lists all responses and then ranks future moves to find the best option. Turochamp is too complex for hardware of the time.	Mind Machines 3e7	50%	Analytic AI
	Logical Thinking	At its core, analytic AI uses logic to process knowledge. Algorithms rely on conditional statements like "IF, THEN, and ELSE." Operators refine the conditions: IF fruit is yellow OR green AND tube-like NOT round, THEN banana.	Mind Machines 5e7	100%	Analytic AI
	Data Processing	As computer power grows, data analysis gains sophistication. Algorithms analyze two types of data. Numbers are counted, compared, and combined into statistics. Non-numbers (images, words, etc.) are organized into classes and ranked.	Mind Machines 1.75e8	100%	Analytic AI Turochamp
	Brute Force	Logic requires stepping through every option before finding the best one—a strategy called brute force. An AI seeking the shortest route between two cities plots and ranks every possible path. This takes time and processing power.	Mind Machines 4e8	50%	Logical Thinking
	Heuristics	Heuristics are algorithmic shortcuts to the "every step" approach of brute force logic. At key points along a route, they could guide an AI to choose the better option. A "good enough" path is mapped out faster.	Mind Machines 1e9	75%	Brute Force
	Fuzzy Logic	An AI can make decisions based on degrees of true, a point between the binary choices of true or false. When dealing with incomplete or large data sets, this method could be used, such as to determine if the weather is hot or cold.	Mind Machines 2.5e9	50%	Heuristics
	Games, Mastered	AI computers surpass humans at games with narrow tasks and fixed rules. In 1997, Deep Blue beats the world chess champion. In 2011, Watson wins the trivia game Jeopardy. In 2016, AlphaGo prevails at the strategy game Go.	Mind Machines 1.5e10	50%	Turochamp ELIZA Therapy Bot
	Expert Systems	An AI can be fed information about a topic, then be programmed to copy the decision-making logic of specialists. This tool can help human experts work more efficiently and also act as a database of knowledge for future users.	Mind Machines 2e10	50%	Games, Mastered
	Context	Logic is only one approach for making sense of the world. AIs struggle with nuances such as sarcasm and other context-based behaviors that humans learn about through life experience and intuition.	Mind Machines 5e10	50%	Data Processing
	Black Box	AI can be considered a 'black box': a system where the input and output are known, but the system's inner workings are not. Without this knowledge, humans may not know if an AI is sentient or simply mimicking human behavior.	Mind Machines 7e10	50%	Context

Generative AI Efficiency

Generative AI has 11 upgrades, increasing the generator efficiency with a total x226,800 multiplier.

Icon	Name	Description	Cost	Efficiency	Requires
	Electric Brain	Each human brain cell receives electrochemical input, triggering an output. Linked together, they can adapt to changes to input in real-time. In 1943, researchers envision an Electric Brain, an AI that mimics how neurons work.	Mind Machines 2e7	50%	Generative AI Imitation Game
	Neural Network	Early brain-model AIs teach themselves simple tasks, like identifying animals. At each node in the network, the algorithm sorts input into a binary output—say, dog or cat. It may fail, but over millions of trials, output improves.	Mind Machines 1e11	1,400%	Electric Brain Stochastic Parrot
	Machine Learning	Neural nets can teach themselves, analyzing results and tweaking them based on the data they are fed. Machine learning takes off as computers surge in processing power, speed, and memory in the early 2000s.	Mind Machines 1.25e11	900%	Generative AI Stochastic Parrot
	Training Sets	Feed an AI loads of data, like every recipe on the internet. It writes tons of new recipes full of errors and nonsense. Using machine learning, the AI adjusts its rules to improve output and, in time, invents a recipe you can eat.	Mind Machines 2e11	200%	Generative AI Stochastic Parrot
	Big Data	Take the mobile phone locations of billions of people, tracked every second. As data sets become global, big data amps up the training input for AIs, which use it to make predictions, such as which roads have the most traffic.	Mind Machines 5e11	100%	Neural Network
	Large Language Models	AIs can take in and process vast language sets with the goal of making predictions: for example, which words follow other words most often? LLM training produces useful, but imperfect, tools for grammar, translation and searching.	Mind Machines 9e11	200%	Training Sets
	GPTs	Generative Pretrained Transformers analyze and generate text. Neural nets like ChatGPT begin writing and conversing with humans at an uncanny level. Based on human-created training sets, they generate code, images, and videos.	Mind Machines 1.5e12	100%	Large Language Models
	Deep Learning	Imagine neural nets with millions of layers and billions of nodes. Data is passed down, node-to-node, to deep, hidden layers. AIs shock us with novel, unexpected output. It is unknown how, exactly, they learn to solve problems.	Mind Machines 3e12	100%	Machine Learning
	Natural Language	Neural net AIs teach themselves to adapt to language that is varied, nuanced, and constantly changing—a huge leap in intelligence. Natural language processors and voice recognition tools make AIs sound more human.	Mind Machines 9e12	100%	GPTs
	Hallucinations	When AIs 'hallucinate', they are making stuff up based on poor quality input. They don't know what they're saying or whether it is wrong. These hallucinations can be prevented by higher quality input and human oversight.	Mind Machines 1.5e13	100%	Natural Language
	AI Evolved	The goal for AI is for them to operate, adapt, and learn on their own. As the nodes of their neural nets improve, they get closer to being able to navigate city streets, explore outer space, and even mimic humans.	Mind Machines 4e13	250%	Deep Learning

AI Tools Efficiency

AI Tools has 12 upgrades, increasing the generator efficiency with a total x31,104 multiplier.

Icon	Name	Description	Cost	Efficiency	Requires
	Personal Assistants	AI is here and in the hands of millions. Phones, watches, and other devices feature assistants that can converse very much like humans. These fast learners improve and personalize their output through use.	Mind Machines 3e14	100%	AI Tools
	Customer Service	Human service reps are trained to follow scripts, which are easily fed to chatbots. Call centers now rely on a blend of the two. Humans step in when the AI detects a red flag or signals it is confused.	Mind Machines 6e14	200%	Personal Assistants
	Therapists	Equipped with medical databases and warm voices, some AIs can help patients cope with trauma. They offer hope and solutions with zero judgment and are available at all hours of every day. As with ELIZA, some humans grow attached.	Mind Machines 9e14	100%	Customer Service
	Translators and Interpreters	AIs can translate spoken and written texts well enough for most uses. The newest leap is real-time interpretation, translating the words as they are spoken. Star Trek's Universal Translator is almost here.	Mind Machines 3e15	100%	Therapists
	AI Avatars	An AI trained on a person's voice, image, and biography can serve as a digital avatar. Users can talk to their favorite fictional characters or can learn about history from an avatar of a holocaust survivor.	Mind Machines 6e15	200%	Translators and Interpreters
	Self-Driving Vehicles	Robotic cars and trucks sense and interact with the world. Where other AIs deal with set rules, self-driving vehicles must navigate unknown obstacles and unpredictability—including human drivers.	Mind Machines 1.25e17	100%	AI Tools
	AI Drones and Carts	Smart robo-carts navigate point-to-point to deliver goods in factories and, increasingly, in towns and cities. Airborne delivery drones contend with fewer obstacles, moving packages over long distances from warehouse to mailbox.	Mind Machines 2e17	200%	Self-Driving Vehicles
	Space Explorers	Far-flung space probes, orbiters, and rovers rely on coded commands from Earth, which can take hours or days to reach them. AI gives them the intelligence and agency to act, think, make decisions, and solve problems on their own.	Mind Machines 6e17	100%	AI Drones and Carts
	Statistical Forecasting	AIs are far better and way faster than humans at analyzing complex data sets and modeling future events. Weather forecasting, financial analysis, and crime prevention are just three of the endless applications.	Mind Machines 3e18	100%	AI Tools AI Avatars
	Smarter Systems	AIs manage complex systems faster and more effectively than humans. For example, they help farmers analyze a wide range of real-time and historic data to determine when, where, and what crops to plant.	Mind Machines 6e18	200%	Statistical Forecasting
	Scientific Discoveries	AIs shine bright in science. They discover exoplanets, black holes, and galaxies. They fold new 3D proteins to advance medicine. They compare whale songs to their behavior, allowing humans to understand their language.	Mind Machines 1e19	100%	Smarter Systems
	Mind Reading	Electrodes record a person's brain signals, and AIs can analyze those signals and reproduce them—in effect, reading minds. The tech has helped paralyzed people walk by controlling brain-muscle commands.	Mind Machines 1.75e19	200%	Scientific Discoveries

Evils and Perils Efficiency

Evils and Perils has 10 upgrades, increasing the generator efficiency with a total x5,184 multiplier.

Icon	Name	Description	Cost	Efficiency	Requires
	Bad Actors	A top fear of AI isn't the tech itself. It's the people who misuse it to enrich and empower themselves and to harm others. Governments, companies, and powerful groups can weaponize AI tools on a global scale.	Mind Machines 1e16	100%	Evils and Perils
	Gullibility	Schemes to steal data, extort money, or hack into systems rely on fooling people. AI deepfakes generate voices, images, and videos that are extremely hard to tell from real ones, making anyone vulnerable to being tricked.	Mind Machines 2e16	100%	Bad Actors
	Human Bias	AI systems trained on internet data reflect humanity, flaws and all. Image generators lean toward showing near unattainable beauty standards, and predictive crime algorithms are based on data skewed by human bias.	Mind Machines 6e16	200%	Gullibility
	Resistance	People rally around fear of AI. They put traffic cones on the hoods of robo-taxis, disabling them. Artists design makeup to block facial recognition. They plant poison pixels in digital works to muddle AI images derived from them.	Mind Machines 9e16	100%	Human Bias
	Weak Laws	Tech often outpaces the slow road to making laws and safeguards. But a wild west of unfettered AI poses new dangers. AIs able to act on their own could transcend our ability to understand and control them.	Mind Machines 9e17	200%	Evils and Perils
	AI Workforce	AIs can make some workers obsolete—drivers, reporters, and coders to name a few. AIs can write, create art, and pass the bar exam. By producing "good enough" results, they will mostly replace entry-level workers.	Mind Machines 1e18	100%	Weak Laws
	Cyberattacks and Warfare	The power of AI can be harnessed to kill people and destroy society. With AI tools, battle drones become stealthy killing machines. Coupled with AI, cyberattacks multiply in speed, power, and disruption.	Mind Machines 2e18	100%	AI Workforce
	Machine Unlearning	How can human bias, errors, and other bad input be removed from trained AI systems? Once integrated, the data becomes a part of the neural net, and it can't be unlearned. The AI would need to be rebuilt from scratch.	Mind Machines 4e19	100%	Evils and Perils Resistance
	Red Teaming	To keep AI systems in check, ethical hackers are hired to stretch and break them. Called red teams, these hackers expose vulnerabilities so they can be patched. But, as AI continuously evolves, new glitches will always appear.	Mind Machines 8e19	200%	Machine Unlearning
	Tough Decisions	How can an AI be taught to make tough decisions with no clear answer? Should robo-taxis always obey traffic laws if doing so can cause a crash? Who would be held accountable for an AI-related accident?	Mind Machines 1e20	200%	Red Teaming

Artificial Intelligence Efficiency

Artificial Intelligence has 40 upgrades, increasing the generator efficiency with a x2.83717e8 payout multiplier, and a x1.85834e9 speed multiplier

Icon	Name	Description	Cost	Payout	Speed	Requires
Upgrades for other generators that also affect this one
	Learning Loops	What sets AI apart? The ability to learn—to evaluate output and improve it. Algorithms with learning loops can include human input, or not. An AI can be trained to identify images by adding tags and flagging errors.	Mind Machines 500,000		4x	Algorithms
	Brute Force	Logic requires stepping through every option before finding the best one—a strategy called brute force. An AI seeking the shortest route between two cities plots and ranks every possible path. This takes time and processing power.	Mind Machines 4e8		0.5x	Logical Thinking
	Stochastic Parrot	AIs often babble, like a parrot. As they mimic their input over and over, they spout random nonsense with no understanding or meaning. Bad or limited input (GIGO) and faulty algorithms are to blame.	Mind Machines 9e9		0.75x	ELIZA Therapy Bot
	Games, Mastered	AI computers surpass humans at games with narrow tasks and fixed rules. In 1997, Deep Blue beats the world chess champion. In 2011, Watson wins the trivia game Jeopardy. In 2016, AlphaGo prevails at the strategy game Go.	Mind Machines 1.5e10		9x	Turochamp ELIZA Therapy Bot
	Expert Systems	An AI can be fed information about a topic, then be programmed to copy the decision-making logic of specialists. This tool can help human experts work more efficiently and also act as a database of knowledge for future users.	Mind Machines 2e10	51x		Games, Mastered
	Context	Logic is only one approach for making sense of the world. AIs struggle with nuances such as sarcasm and other context-based behaviors that humans learn about through life experience and intuition.	Mind Machines 5e10		0.75x	Data Processing
	Black Box	AI can be considered a 'black box': a system where the input and output are known, but the system's inner workings are not. Without this knowledge, humans may not know if an AI is sentient or simply mimicking human behavior.	Mind Machines 7e10		0.75x	Context
	GPTs	Generative Pretrained Transformers analyze and generate text. Neural nets like ChatGPT begin writing and conversing with humans at an uncanny level. Based on human-created training sets, they generate code, images, and videos.	Mind Machines 1.5e12		10x	Large Language Models
	Deep Learning	Imagine neural nets with millions of layers and billions of nodes. Data is passed down, node-to-node, to deep, hidden layers. AIs shock us with novel, unexpected output. It is unknown how, exactly, they learn to solve problems.	Mind Machines 3e12		10x	Machine Learning
	AI Evolved	The goal for AI is for them to operate, adapt, and learn on their own. As the nodes of their neural nets improve, they get closer to being able to navigate city streets, explore outer space, and even mimic humans.	Mind Machines 4e13	51x		Deep Learning
	Personal Assistants	AI is here and in the hands of millions. Phones, watches, and other devices feature assistants that can converse very much like humans. These fast learners improve and personalize their output through use.	Mind Machines 3e14	Automate		AI Tools
	Customer Service	Human service reps are trained to follow scripts, which are easily fed to chatbots. Call centers now rely on a blend of the two. Humans step in when the AI detects a red flag or signals it is confused.	Mind Machines 6e14		2x	Personal Assistants
	Therapists	Equipped with medical databases and warm voices, some AIs can help patients cope with trauma. They offer hope and solutions with zero judgment and are available at all hours of every day. As with ELIZA, some humans grow attached.	Mind Machines 9e14		2x	Customer Service
	Translators and Interpreters	AIs can translate spoken and written texts well enough for most uses. The newest leap is real-time interpretation, translating the words as they are spoken. Star Trek's Universal Translator is almost here.	Mind Machines 3e15	2x		Therapists
	AI Avatars	An AI trained on a person's voice, image, and biography can serve as a digital avatar. Users can talk to their favorite fictional characters or can learn about history from an avatar of a holocaust survivor.	Mind Machines 6e15	2x		Translators and Interpreters
	Bad Actors	A top fear of AI isn't the tech itself. It's the people who misuse it to enrich and empower themselves and to harm others. Governments, companies, and powerful groups can weaponize AI tools on a global scale.	Mind Machines 1e16		0.75x	Evils and Perils
	Gullibility	Schemes to steal data, extort money, or hack into systems rely on fooling people. AI deepfakes generate voices, images, and videos that are extremely hard to tell from real ones, making anyone vulnerable to being tricked.	Mind Machines 2e16		0.75x	Bad Actors
	Human Bias	AI systems trained on internet data reflect humanity, flaws and all. Image generators lean toward showing near unattainable beauty standards, and predictive crime algorithms are based on data skewed by human bias.	Mind Machines 6e16		0.75x	Gullibility
	Resistance	People rally around fear of AI. They put traffic cones on the hoods of robo-taxis, disabling them. Artists design makeup to block facial recognition. They plant poison pixels in digital works to muddle AI images derived from them.	Mind Machines 9e16		0.75x	Human Bias
	Self-Driving Vehicles	Robotic cars and trucks sense and interact with the world. Where other AIs deal with set rules, self-driving vehicles must navigate unknown obstacles and unpredictability—including human drivers.	Mind Machines 1.25e17		2x	AI Tools
	AI Drones and Carts	Smart robo-carts navigate point-to-point to deliver goods in factories and, increasingly, in towns and cities. Airborne delivery drones contend with fewer obstacles, moving packages over long distances from warehouse to mailbox.	Mind Machines 2e17		2x	Self-Driving Vehicles
	Space Explorers	Far-flung space probes, orbiters, and rovers rely on coded commands from Earth, which can take hours or days to reach them. AI gives them the intelligence and agency to act, think, make decisions, and solve problems on their own.	Mind Machines 6e17		5x	AI Drones and Carts
	Weak Laws	Tech often outpaces the slow road to making laws and safeguards. But a wild west of unfettered AI poses new dangers. AIs able to act on their own could transcend our ability to understand and control them.	Mind Machines 9e17		0.5x	Evils and Perils
	AI Workforce	AIs can make some workers obsolete—drivers, reporters, and coders to name a few. AIs can write, create art, and pass the bar exam. By producing "good enough" results, they will mostly replace entry-level workers.	Mind Machines 1e18		0.5x	Weak Laws
	Cyberattacks and Warfare	The power of AI can be harnessed to kill people and destroy society. With AI tools, battle drones become stealthy killing machines. Coupled with AI, cyberattacks multiply in speed, power, and disruption.	Mind Machines 2e18		0.5x	AI Workforce
	Statistical Forecasting	AIs are far better and way faster than humans at analyzing complex data sets and modeling future events. Weather forecasting, financial analysis, and crime prevention are just three of the endless applications.	Mind Machines 3e18		10x	AI Tools AI Avatars
	Smarter Systems	AIs manage complex systems faster and more effectively than humans. For example, they help farmers analyze a wide range of real-time and historic data to determine when, where, and what crops to plant.	Mind Machines 6e18	6x		Statistical Forecasting
	Scientific Discoveries	AIs shine bright in science. They discover exoplanets, black holes, and galaxies. They fold new 3D proteins to advance medicine. They compare whale songs to their behavior, allowing humans to understand their language.	Mind Machines 1e19		11x	Smarter Systems
	Mind Reading	Electrodes record a person's brain signals, and AIs can analyze those signals and reproduce them—in effect, reading minds. The tech has helped paralyzed people walk by controlling brain-muscle commands.	Mind Machines 1.75e19		10x	Scientific Discoveries
	Machine Unlearning	How can human bias, errors, and other bad input be removed from trained AI systems? Once integrated, the data becomes a part of the neural net, and it can't be unlearned. The AI would need to be rebuilt from scratch.	Mind Machines 4e19		0.75x	Evils and Perils Resistance
	Red Teaming	To keep AI systems in check, ethical hackers are hired to stretch and break them. Called red teams, these hackers expose vulnerabilities so they can be patched. But, as AI continuously evolves, new glitches will always appear.	Mind Machines 8e19		0.75x	Machine Unlearning
	Tough Decisions	How can an AI be taught to make tough decisions with no clear answer? Should robo-taxis always obey traffic laws if doing so can cause a crash? Who would be held accountable for an AI-related accident?	Mind Machines 1e20		2x	Red Teaming
Upgrades that only affect this generator
	Future AI	Amara's law states: We overestimate the effect of a technology in the short run and underestimate it in the long run.	Mind Machines 1.5e20		50x	Mind Reading Tough Decisions
	Quantum Computing	These computers replace circuits with subatomic particles and qubits, which process multiple dimensions with immense speed, power, and volume. They'll enable AI to make exponentially more complex connections across diverse data sets.	Mind Machines 2e20		2x	Future AI
	Biocomputers	Brain cells act like circuits. Some scientists build on this by creating electric brains from human neurons. These biocomputers use less energy, but raise questions about ethical use of biomatter outside of a medical context.	Mind Machines 3.5e20		2.5x	Quantum Computing
	Uploaded Mind	Is the human mind a type of computer programmed to think? Could it be copied, transferred, uploaded to a server after death? Would an artificial copy of a mind be any different from the original?	Mind Machines 7e20		2.5x	Biocomputers
	Philosophical Zombies	Skeptics insist AIs will always be 'philosophical zombies': humanlike in appearance, but devoid of interests, opinions, and life. These 'zombies' can still evoke genuine emotion in humans despite not having their own.	Mind Machines 1.5e21	3x		Future AI Uploaded Mind
	Morality AI	Can AIs learn human values? If so, whose values? Some humans might prioritize quantity over quality, others may suggest a human life is not as valuable as a product. An AI's vision of right and wrong depend on the people training it.	Mind Machines 4e21	3x		Future AI Philosophical Zombies
	Sentience	AIs are able to do things they're not trained to do. But are they aware of their existence? Is imitating sentience the same as being sentient? Are humans sentient, or are their brains programmed to think they are?	Mind Machines 1e22	5x		Future AI Morality AI
	Artificial General Intelligence	The dream of AGI is to equal or surpass humans in all intelligences, whether it be logic, creativity, or personal and emotional skills. Impossible, say skeptics. Inevitable, say believers.	Mind Machines 1.5e22	101x		Sentience

Tech Tree

Trivia

• In the beta, Semblance's opening quote was "Who am I? What am I? Am I really that different from you?".
• This is the first Season 2 exploration that has an alternative currency.
• The Self-Driving Vehicles node bears a striking resemblance to the Self-Driving Car node from the Primary Simulation.
• The tech tree looks like Semblance.