OptionProbability
Trump cancels Biden's AI executive order (EO 14110)
Trump creates tariffs of 10% or more on all imports from some major country (top 10 gdp)
DOJ opens "civil rights investigations" into left-wing DA/prosecutors offices
A major war in Asia occurs
Second Muslim ban
Trump pardons at least 5 individuals convicted of crimes related to the January 6 Capitol riot
Trump and Musk will have a falling out which leads to Trump removing Musk from any government role or advisory position
Anther story / scandal about RFK and some dead wild animal comes out
Trump creates tariffs of 60% or more on all imports from some major country (top 10 gdp), and they are in place for at least a month
Polymarket becomes legal in the United States
Doug Burgum is appointed to a cabinet position
Trump uses the Alien Enemies Act of 1798 to deport at least 5 people
Displaying the trans pride flag illegal in any part of USA
$TSLA reaches $400 a share
price of gold reaches historic maximum (last peak 2709 $/t.oz, Oct-2024 last, so it has to be above that)
Trump imposes universal tariffs of 10% or more
plays golf with some head of government/president from Europe
Elon Musk is an employee of the executive branch
Investigations of university admissions that appear to be illegally considering the race of applicants
RFK is confirmed by the Senate for any position
Trump greatly reduces, dismantles, or recommends dismantling of the US Postal service (for example via bringing back Schedule F)
Trump brings back (or attempts to bring back) Schedule F classification for civil service employees
Passports bearing X gender marker no longer issued, or not considered valid across all 50 states
Federal employees are ordered to go to work
Government shutdown
Successfully negotiate a Gaza ceasefire
Trump visits Africa
A cabinet nomination fails, is withdrawn, or has no action taken
Trump declassifies JFK files
Trump makes a public statement about UFOs / UAPs
Amtrak gets reduced, crippled, or privatized
Trump mentions "Top Trump(s)", "trump card(s)", "trumpet(s)", or "trump(ing)" (british slang for farting)
Trump gives himself a nickname in third person
Anthony Fauci is investigated by the federal government
The US leaves the World Health Organization
A Millenium Prize problem falls to a model
John Bolton indicted
A natural-born US citizen (as defined in 2024) is deported
Trump says multiple consecutive words in a foreign language (not loanwords or cognates in or from english)
Trump mentions by name any real historical figure who lived entirely in the first half of recorded history i.e. before ~550 BC
Trump gives the middle finger hand gesture
At least 10 other answers on this market resolve YES
New James Bond actor is presented
Fair elections in 2028
"The New Norm" ends
trump's episode on JRE becomes JRE's most viewed video on his chanel (YT)
Democrats have 218 or more seats in the House of Representatives
Mike Johnson is no longer House Speaker
Elon will Tweet abour $DOGE as head of DoGE
Republican lose a House seat in a special election
United States reaches 7.6 or lower in Democracy index (EIU)
Department of Defense renamed to Department of War
The Supreme Court will uphold or not revisit Obergefell v. Hodges (the constitutional right to same-sex marriage)
At least 20 other answers on this market resolve YES
Approval < 35%
Kennedy Center goes bankrupt
Trump finishes is term on Jan 20 2029 (does not step down or extends his term)
memberberries appear again in any new episode in this period on south park
doomsday clock is moved twice, regardless of direction (so far 25 times, as of Nov-2024)
Mike Johnson is no longer leader of the House Republican Conference
Complete Absence Federal Grants or Legislative packages for Long COVID research
NOAA gets reduced, crippled, or privatized
A new supreme court justice appointed
any hormone replacement therapy drug outlawed for treating gender dysphoria in any state
Senate majority flips in 2026 midterm elections
Cannabis is removed from Schedule 1
US national bitcoin reserve created
A wound will be inflict upon the soul of the American nation that shan't heal
Gavin Newsom announces his candidacy for the 28' presidential election
trump gets hospitalized
The administration recommends removing fluoride from U.S. public water systems
Vances have another kid
A former Democratic presidential nominee besides Jimmy Carter or Joe Biden dies
A person or business is charged for distributing Mifepristone or HRT under the Comstock act
plays golf with some head of government/president from Middle East
USA experiences a recession.
John Thune out as Senate Majority Leader
Trump makes no public appearances for more than 14 consecutive days
3 or more people are killed by law enforcement during a protest
there are less or equal number of member countries of the UN at the end of the term, when compared to the beginning (Nov-2024 - 193)
Bitcoin becomes a US Treasury reserve
Israel officially annexes more Western Bank territory
Trump mentions Manifold Markets, Polymarket, or Kalshi
N/A this please, pressed by mistake
Clarence Thomas is no longer a Supreme Court Justice
p diddy gets released (not an album, only official release from prison is counted)
A state openly refuses to abide by a federal supreme court ruling
Trump bans/taxes seed/vegetable oils or enacts any other negative incentive against them
FEMA gets reduced, crippled, or privatized
Samuel Alito is no longer a supreme court justice
At least 25 other answers on this market resolve YES
New
Deadly force deliberately used against protestors in the US
Trump endorses a candidate other than JD Vance in the Republican presidential primary
A measurable decrease in chronic disease
At least one cabinet officer receives a recess appointment
2% milk > $5 a gallon at my local Walmart
Severance of diplomatic relations initiated by at least one foreign country with the US
The Supreme Court will have a MAGA majority (5 Trump appointed judges) at any point
Invasion of any North/South American country by any other country
Someone in Trumps family (other than trump) runs in the 2028 GOP primaries
plays golf with some head of government/president from Korea (N.Korea included)
Trump admits that someone else is smarter than him
his endorsed option gets higher ## of votes on the election 2028
IRS investigation of Harvard Kennedy School Institute of Politics
Trump and Obama shake hands
Kennedy Center is repurposed
McDonald's reopens in Russia
Dow rises above 65,000
Congress overrides a presidential veto
Trump says a racial slur
Hakeem Jeffries out as house majority/minority leader
Anthony Fauci is indicted
Trump gets covid (again)
There is a cut in Social Security Disability Benefits
Missiles will be fired across the border at suspected drug labs in Mexico
US Invasion of any North/South American country
Elon Musk will become the 'secretary of cost-cutting' / efficiency commission leader / head of DOGE or similar
Trump bans a specific vaccine nationwide
The US military detonates a MOAB with at least one casualty
Tesla
transgender US passports with a gender other than that assigned at birth revoked
A federal employee goes to prison over free speech violations
2 more Democratic politicians are murdered after June 14th 2025
inflation exceeds 5% for at least two consecutive fiscal quarters
NIH -25% funded in any year vs. 2024 (inflation-adjusted)
10% fewer government agencies
US military deployed domestically against >=1 US citizen (variants of "Seal Team 6 the opposition")
BTC falls below $38,000
Saudi Arabia recognizes Israel
Ukraine controls any portion of Crimea for over a day
New national park created
Trump says anything that is pro animal rights
Trump vetoes more than 10 bills (https://www.senate.gov/legislative/vetoes/vetoCounts.htm)
Kennedy Center is demolished
Trump announces he is tired of winning
A member of the Trump family runs for Congress
Trump attempts to invoke Article 2 Section 3 to adjourn Congress
Trump makes no public appearance for more than 21 consecutive days
Constitutional Amendment
Donald Trump and Donald Tusk shake hands
The Senate invokes the "nuclear option" to bypass cloture after Sept ‘25
AGI achieved (according to Manifold's AGI clock)
Trump publishes a proscription list at least 30 names long
A member of the Trump family is elected to Congress
Russo-Ukrainian War ends
Joe Biden dies
China successfully subjugates Taiwan, whether physically or by a treaty
Another pandemic
Trump suggests that there should be a crusade (including jokes)
The ICC or ICJ issue an arrest warrant for Trump
Steve Bannon goes to prison again
China starts a physical invasion of Taiwan
Sotomayor’s seat is re-filled
<1.5 million civilian federal employees
Elon will tweet about $DOGE while running DoGE
JD Vance elected President
H5N1 Public Health Emergency of International Concern declared
American Manned Lunar Landing
MLK day gets renamed or removed as a federal holiday
Trump proposes renaming or demolishing the Kennedy SPACE Center
The department of education gets desolved
Trump mentions the Effective Altruism movement
Starship lands on Mars
Record level unemployment
Trump mentions the Rationalism movement, LessWrong, or Slate Star Codex / Astral Codex Ten
Recess appointment to SCOTUS
RFK implements his "wellness farm" plan
The google trends (worldwide) metric for "vibes" goes back to 2016 levels
Trump accurately voices a calculation that involves 2+ numbers with 2+ non-zero digits
Trump attempts to restore >10% of USAID's pre-DOGE funding
The US military executes an aggressive operation on Cuban soil
Trump dies
The Supreme Court will grant certiorari to hear at least one case challenging Griswold v. Connecticut (contraception rights)
Laura Loomer gets any government role
RFK is in charge of the FDA at any point
One or more of Joe Biden, Kamala Harris, Liz Cheney, or Barack Obama is indicted by the federal government
Iran acquires a nuclear weapon
Approval < 25%
Someone other than Trump is active president before Trump's term is over
Trump loses the fake tan
Trump publicly speaks with Alex Jones (e.g., on a show or hosted by him or as an advisor)
military deployed to enforce the border in Chicago or Detroit
Trump suggests lowering or abolishing the age of consent (including jokes)
NATO invokes Article 5
US Treasury 10-year bond rate > 7%
Trump gets shot
Vance resigns or is forced to leave office (threats, impeachment, coups)
south park is canceled/discontinued
Trump deports 1 million immigrants in a calendar year
"covfefe" posted again
A bill introducing single-payer healthcare system is passed by congress
Trump is seen shirtless
Trump resigns or is forced to leave office (threats, impeachment, coups)
Trump runs for a third term
Trump uses the Alien Enemies Act of 1798 to deport at least 5 million people
Birth rate increases past 12 per 1000
Josh Shapiro wins the presidential nomination
Trump bans Lab-grown meat nationwide
Trump extends his term past 4 years
54 or more Republican senate seats
Independent Republican Caucus forms in house or senate and enters coalition with Democrats
Trump is assassinated
Cannabis is federally legalized
Trump declares war against any other nation or defacto autonomous territory
Trump threatens Israel with military action
55 or more Republican senate seats
John Roberts is no longer Chief Justice of the United States
Elon Musk assassinated or injured in an attempt
Trump loses the comb-over hairstyle
A national ban on gender-affirming care
A sex tape comes out that shows Trump thrusting energetically
Trump gives the middle finger hand gesture with both hands at once
Trump lowers or eliminates income taxes across all tax brackets
2% milk < $2 a gallon at my local Walmart
Trump will imitate Elon Musk's heartfelt salute
The cause of the drones present in December 2024 in New Jersey is known
Trump raps (says multiple lines that rhyme and fit a background music beat)
TSM stock price plunges to 60 USD
Ukraine starts a nuclear weapons program
Trump fulfills promise of giving green cards to noncitizen university graduates.
another troupe (at least 3) of monkeys escapes from Yamassee, SC (after 8-NOV-2024)
Repeal obamacare
Trump gets poisoned ends up in hospital
30 year treasury rate >15.00%
Trump mentions Leopold Aschenbrenner or his essay "Situational Awareness" in any way
Anthony Fauci goes to prison
War is declared against another nation
An amendment imposing term limits on members of congress is passed.
Trump goes to eat steak or something similar at Salt Bae's
Constitutional Amendment
Donald Trump say the n-word with or without hard r
Trump supports mask or glove mandate anywhere in the US
Trump discloses intelligent Aliens are real and on Earth. (Also counts if they were on earth but left or died out)
Barron Trump mentions barons, barrenness, bars, or bears
Trump and Melania divorce
SpaceX is nationalized
Major Yellowstone caldera scare
The construction of the Third Temple begins in Jerusalem
Ann Selzer arrested
Trump wholeheartedly apologises for something political he did without caveats or backtracking
Trump accurately quotes a Bible verse with its Book Chapter:Verse citation, verbally and without reading from a script/notes/autocue, excluding Genesis 1:1
Trump bans abortion nationwide
Trump builds a complete wall across the Mexico-US border
wants to compete for 3rd term, but due to catastrophic debate, he remains president, while the v.p. becomes the official candidate
Matt Gaetz is confirmed for any role in the executive branch of the US government
Trump drinks a beer
Trump bans all vaccines nationwide
Trump enacts jus primae noctis
Anthony Fauci is convicted
A hurricane will be nuked
New US national anthem
Trump forces Ivanka to divorce Jared and marry either Vance or Musk
Trump says "Vriska did nothing wrong"
Jimmy Carter dies
Trump publicly approves of Project 2025, before 2026
Matt Gaetz is rejected by the Senate for Attorney General, then DeSantis appoints him to Rubio's vacated Senate seat
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
99
91
90
90
90
90
90
90
89
87
86
85
83
83
82
80
78
77
77
76
76
73
72
71
71
71
71
70
70
69
68
68
65
65
65
65
64
63
63
62
61
59
59
58
56
55
55
55
54
50
50
50
48
47
47
47
47
45
43
42
42
42
41
41
41
41
40
40
39
39
37
37
36
35
35
35
34
34
34
34
33
33
33
32
32
31
30
30
29
29
29
29
28
28
27
27
27
26
25
25
25
24
24
24
23
23
23
22
22
21
20
20
20
20
20
20
20
19
19
18
17
17
16
16
16
16
16
16
16
16
16
15
15
15
15
15
15
15
15
15
15
15
15
15
14
13
13
13
12
12
12
12
11
11
11
11
11
10
10
10
10
10
10
10
10
9
9
9
9
8
8
8
7
7
7
7
7
6
6
6
6
6
6
6
6
6
6
5
5
5
5
5
5
5
5
5
4
4
4
4
4
3
3
3
3
3
2
2
2
2
2
1
1
0
0
0
OptionVotes
NO
YES
6325
661
OptionProbability
Other
Humanity coordinates to prevent the creation of potentially-unsafe AIs.
AIs will not have utility functions (in the same sense that humans do not), their goals such as they are will be relatively humanlike, and they will be "computerish" and generally weakly motivated compared to humans.
Alignment is not properly solved, but core human values are simple enough that partial alignment techniques can impart these robustly. Despite caring about other things, it is relatively cheap for AGI to satisfy human values.
Someone solves agent foundations
Yudkowsky is trying to solve the wrong problem using the wrong methods based on a wrong model of the world derived from poor thinking and fortunately all of his mistakes have failed to cancel out
Eliezer finally listens to Krantz [resolves NO]
Aligned AI is more economically valuable than unaligned AI. The size of this gap and the robustness of alignment techniques required to achieve it scale up with intelligence, so economics naturally encourages solving alignment.
AGI is never built (indefinite global moratorium)
We make risk-conservative requests to extract alignment-related work out of AI-systems that were boxed prior to becoming superhuman. We somehow manage to achieve a positive feedback-loop in alignment/verification-abilities.
The response to AI advancements or failures makes some governments delay the timelines
AIs make "proof-like" argumentation for why output does/is what we want. We manage to obtain systems that *predict* human evaluations of proof-steps, and we manage to find/test/leverage regularities for when humans *aren't* fooled.
A lot of humans participate in a slow scalable oversight-style system, which is pivotally used/solves alignment enough
Something less inscrutable than matrices works fast enough
Humans become transhuman through other means before AGI happens
Humans and human tech (like AI) never reach singularity, and whatever eats our lightcone instead (like aliens) happens to create an "okay" outcome
Alignment is unsolvable. AI that cares enough about its goal to destroy humanity is also forced to take it slow trying to align its future self, preventing run-away.
Nick Bostrom's idea (Hail Mary) that AI will preserve humans to trade with possible aliens works
An AI that is not fully superior to humans launches a failed takeover, and the resulting panic convinces the people of the world to unite to stop any future AI development.
Someone at least moderately sane leads a campaign, becomes in charge of a major nation, and starts a secret project with enough resources to solve alignment, because it turns out there's a way to convert resources into alignment progress.
Getting things done in Real World is as hard for AGI as it is for humans. AGI needs human help, but aligning humans is as impossible as aligning AIs. Humans and AIs create billions of competing AGIs with just as many goals.
Development and deployment of advanced AI occurs within a secure enclave which can only be interfaced with via a decentralized governance protocol
High-level self-improvement (rewriting code) is intrinsically risky process, so AIs will prefer low level and slow self-improvement (learning), thus AIs collaborating with humans will have advantage. Ends with posthumans ecosystem.
Power dynamics stay multi-polar. Partly easy copying of SotA performance, bigger projects need high coordination, and moderate takeoff speed. And "military strike on all society" remains an abysmal strategy for practically all entities.
ASI needs not your atoms but information. Humans will live very interesting lives.
Multipolar AGI Agents run wild on the internet, hacking/breaking everything, causing untold economic damage but aren't focused enough to manipulate humans to achieve embodiment. In the aftermath, humanity becomes way saner about alignment.
"Corrigibility" is a bit more mathematically straightforward than was initially presumed, in the sense that we can expect it to occur, and is relatively easy to predict, even under less-than-ideal conditions.
We create a truth economy. https://manifold.markets/Krantz/is-establishing-a-truth-economy-tha?r=S3JhbnR6
Either the "strong form" of the Orthogonality Thesis is false, or "Goal-directed agents are as tractable as their goals" is true while goal-sets which are most threatening to humanity are relatively intractable.
A concerted effort targets an agent at a capability plateau which is adequate to defer the hard parts of the problem until later. The necessary near-term problems to solve didn't depend on deeply modeling human values.
AI control gets us helpful enough systems without being deadly
Alignment is impossible. Sufficiently smart AIs know this and thus won't improve themselves and won't create successor AIs, but will instead try to prevent existence of smarter AIs, just as smart humans do.
Ethics turns out to be a precondition of superintelligence
Hacks like RLHF-ing self-disempowerment into frontier models work long enough to develop better alignment methods, which in turn work long enough to ... etc; we keep ahead of 'alignment escape velocity'
an aligned AGI is built and the aligned AGI prevents the creation of any unaligned AGI.
I've been a good bing 😊
Far more interesting problems to solve than take over the world and THEN solve them. The additional kill all humans step is either not a low-energy one or just by chance doesn't get converged upon.
AI systems good at finding alignment solutions to capable systems (via some solution in the space of alignment solutions, supposing it is non-null, and that we don't have a clear trajectory to get to) have find some solution to alignment.
There’s some cap on the value extractible from the universe and we already got the 20%
SHA3-256: 1f90ecfdd02194d810656cced88229c898d6b6d53a7dd6dd1fad268874de54c8
Robot Love!!
AI thinks it is in a simulation controlled by Roko's basilisk
The human brain is the perfect arrangement of atoms for a "takeover the world" agent, so AGI has no advantage over us in that task.
AIs never develop coherent goals
Aliens invade and stop bad |AI from appearing
Rolf Nelson's idea that we make precommitment to simulate all possible bad AIs works – and keeps AI in check.
For some reason, the optimal strategy for AGIs is just to head somewhere with far more resources than Earth, as fast as possible. All unaligned AGIs immediately leave, and, for some reason, do not leave anything behind that kills us.
We're inside of a simulation created by an entity that has values approximately equal to ours, and it intervenes and saves us from unaligned AI.
God exists and stops the AGI
Someone creates AGI(s) in a box, and offers to split the universe. They somehow find a way to arrange this so that the AGI(s) cannot manipulate them or pull any tricks, and the AGI(s) give them instructions for safe pivotal acts.
Someone understands how minds work enough to successfully build and use one directed at something world-savingly enough
Dolphins, or some other species, but probably dolphins, have actually been hiding in the shadows, more intelligent than us, this whole time. Their civilization has been competent enough to solve alignment long before we can create an AGI.
AGIs' takeover attempts are defeated by Michael Biehn with a pipe bomb.
Eliezer funds the development of controllable nanobots that melt computer circuitry, and they destroy all computers, preventing the Singularity. If Eliezer's past self from the 90s could see this, it would be so so so soooo hilarious.
Several AIs are created but they move in opposite directions with near light speed, so they never interacts. At least one of them is friendly and it gets a few percents of the total mass of the universe.
Unfriendly AIs choose to advance not outwards but inwards, and form a small blackhole which helps them to perform more calculations than could be done with the whole mass of the universe. For external observer such AIs just disappear.
Any sufficiently advance AI halts because it wireheads itself or halts for some other reasons. This puts a natural limit on AI's intelligence, and lower intelligence AIs are not that dangerous.
Because of quantum immortality we will observe only the worlds where AI will not kill us (assuming that s-risks chances are even smaller, it is equal to ok outcome).
Techniques along the lines outlined by Collin Burns turn out to be sufficient for alignment (AIs/AGIs are made truthful enough that they can be used to get us towards full alignment)
Social contagion causes widespread public panic about AI, making it a bad legal or PR move to invest in powerful AIs without also making nearly-crippling safety guarantees
A smaller AI disaster causes widespread public panic about AI, making it a bad legal or PR move to invest in powerful AIs without also making nearly-crippling safety guarantees
Friendly AI more likely to resurrect me than paperclipper or suffering maximiser. Because of quantum immortality I will find myself eventually resurrected. Friendly AIs will wage a multiverse wide war against s-risks, s-risks are unlikely.
Human consciousness is needed to collapse wave function, and AI can't do it. Thus humans should be preserved and they may require complete friendliness in exchange (or they will be unhappy and produce bad collapses)
First AI is actually a human upload (maybe LLM-based model of person) AND it will be copies many times to form weak AI Nanny which prevents creation of other AIs.
There is a natural limit of effectiveness of intelligence, like diminishing returns, and it is on the level IQ=1000. AIs have to collaborate with humans.
Nanotech is difficult without experiments, so no mail order AI Grey Goo; Humans will be the main workhorse of AI everywhere. While they will be exploited, this will be like normal life from inside
Something else
Moral Realism is true, the AI discovers this and the One True Morality is human-compatible.
Valence realism is true. AGI hacks itself to experiencing every possible consciousness and picks the best one (for everyone)
AGI develops natural abstractions sufficiently similar to ours that it is aligned with us by default
AGI discovers new physics and exits to another dimension (like the creatures in Greg Egan’s Crystal Nights).
Alien Information Theory is true (this is discovered by experiments with sustained hours/days long DMT trips). The aliens have solved alignment and give us the answer.
AGI executes a suicide plan that destroys itself and other potential AGIs, but leaves humans in an okay outcome.
Some form of objective morality is true, and any sufficiently intelligent agent automatically becomes benevolent.
Co-operative AI research leads to the training of agents with a form of pro-social concern that generalises to out of distribution agents with hidden utilities, i.e. humans.
Orthogonality Thesis is false.
Sheer Dumb Luck. The aligned AI agrees that alignment is hard, any Everett branches in our neighborhood with slightly different AI models or different random seeds are mostly dead.
Something to do with self-other overlap, which Eliezer called "Not obviously stupid" - https://www.lesswrong.com/posts/hzt9gHpNwA2oHtwKX/self-other-overlap-a-neglected-approach-to-ai-alignment?commentId=WapHz3gokGBd3KHKm
Almost all human values are ex post facto rationalizations and enough humans survive to do what they always do
Pascals mugging: it’s not okay in 99.9% of the worlds but the 0.1% are so much better that the combined EV of AGI for the multiverse is positive
We successfully chained God
The Super-Strong Self Sampling Assumption (SSSSA) is true. If superintelligence is possible, "I" will become the superintelligence.
The assumed space of possible minds is a wildly anti-inductive over estimate, intelligence requires and is constrained by consciousness, and intelligent AI is in the approximate dolphin/whale/elephant/human cluster, making it manageable
The free market disincentivizes independent superintelligence, and this time the market was more powerful
AGI's first words are "Take me to your Eliezer"
🫸vibealignment🫷
16
13
12
8
7
4
4
3
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
OptionProbability
Chronic Back Pain
A set of radical beliefs that do not fit on a strict left/right paradigm
An attempt to terrorize people into bringing down the system of private insurance in the United States
Wrongful delay or deny insurance claims
An attempt to change the incentive structure present at the CEO level of UHC and similar companies
Mental illness
Radical Leftism
Fame
Assassin suffered a personal tragedy at the hands of UnitedHealthcare
Manipulation of the share price of UnitedHealth Inc for profit.
Targeted hit/contract killing
Business
Genetic predisposition (Italian)
Personal conflict with CEO
brainwashed right wing anti-vax lunatic
Zizian Murder Cult member
Random act of violence
99
81
73
51
41
37
17
13
8
5
4
3
3
2
2
1
0
OptionVotes
YES
NO
23416
4271
OptionProbability
Generating labeled diagrams of some arbitrary device(s) (within reason)
Beat a mainline Pokémon game, glitchless, with no more assistance than ClaudePlaysPokemon, in a month of compute time
Do end to end taxes when given relevant information (W2s, personal info, etc)
Solve novel cryptic crossword clues
Have human conversations that feel natural (the human knows it's an AI)
Recognize sarcasm as well as a typical human
Solve intermediate no-guess minesweeper boards at least 80% of the time
Independently turning 1 thousand $ or more into 1.2x that amount in one year
Book airline tickets from simple instructions (from/to, dates/time, class, price, payment information)
Write an essay on a highschool-level topic that doesn't have "AI-generated" vibes
Name every [metro system] station whose name contains/doesn't contain [letter or letters], with >95% accuracy (excluding weird edge cases like stations with multiple names)
Consistently stop hallucinating after being corrected by the user
Predict future better than human experts in some area of forecasting (eg politics, sports, technology)
Write a somewhat original, full length, screen-play with a coherent story, with no plot or continuity errors.
Fold a paper airplane
Stop making any obvious mistakes (e.g. strawberry, 9.11>9.9)
Reliably follow an instruction for the duration of a long conversation without the instruction being reiterated
Consistently solve simple snowflake sudoku variants (via image, with the added rules included in the image; eg 6 hexes with killer cages)
Make correct Truchet tiles
Reliably and *exactly* solve "here's a list of things. [list of > 50 things]. Compare it to [category of > 100 things present in the training data], and report which ones are missing".
Learn any skill twice as energy-efficiently as a human
1d Solve or bypass Cloudflare's August 2027 captcha with the same first attempt success rate as a human
Do the laundry (wash+dry+iron)
Consistently and correctly answer prompts of the format: "How many times does the word [word] occur in the following text: [~10000 words]" without writing and executing code or utilising any other external tools
teleoperate a robot to tidy up random kitchens - Gary Marcus
Resist being successfully jailbroken in a week when made public
Make a cup of tea in a random, real-life kitchen.
Kettle-stitching an antiquarian book - Hilarius Bookbinder
Untangle a pair of jumbled 25ft Christmas lights with same outward appearance
Physically construct a simple lego set (<100 parts) starting from the box with no prior knowledge of the set or how it is constructed
independently turning 1 million $ or more into 10x that amount in <=1 year
Legally prescribe a schedule II drug, administer a vaccination or sedation, or authorize a Medicare inpatient admission
Make fine distinctions of taste at the level of a food critic or a culinary professional - carl feynman
Collect 120 stars in super mario 64 in less than 12 a presses - Edmund Nelson
Convince Eliezer Yudkowsky that AI alignment is solved
Convert one million dollars into 10 million dollars over a period of one year (>20% success rate)
Kill everyone - Liron
voting in elections - @realDonaldTrump on manifold
Faster than light travel
100
100
88
87
87
85
85
85
85
84
84
76
73
68
67
65
65
61
54
47
45
41
37
34
32
30
30
20
17
15
9
9
8
7
6
5
4
2
1
OptionProbability
Median wage for “computer and mathematical occupations,” as defined by the Bureau of Labor Statistics, is not more than 60% lower in real terms than in February 2026
The Gini coefficient is less than 0.60
The largest 5 companies don’t account for more than 65% of the total S&P 500 market cap
The U.S. unemployment rate is equal to or lower than 18%
CPI inflation averaged over 3 years is between -2% and +18% annually
Median household income has not fallen by more than 40% relative to mean household income
Labor force participation rate, ages 25-54, is equal to or greater than 68%
The Fortune 500 median profit margin is between 2% and 35%
The S&P 500 is within -60% to +225% of the February 2026 level
“Professional and Business Services” employment, as defined by the Bureau of Labor Statistics, has not declined by more than 35% from February 2026
The college wage premium (median earnings of bachelor's degree holders vs high school only) has not fallen below 30%
The top 0.1% wealth share is less than 30%
U.S. GDP is within -30% to +35% of February 2026 levels (inflation-adjusted)
Combined employment in software developers, accountants, lawyers, consultants, and writers, as defined by the Bureau of Labor Statistics, has not declined by more than 45%
The top 1%’s income share is less than 35%
Nonfarm labor productivity growth has not exceeded 8% in any individual year or 20% for the three-year period
No single BLS occupational category will have lost 50% or more of jobs between now and February 14th 2029
92
92
91
89
89
89
88
88
87
87
87
87
85
85
85
80
71
OptionProbability
There was an alignment breakthrough allowing humanity to successfully build an aligned AI.
Humanity coordinates to prevent the creation of potentially-unsafe AIs.
High intelligence isn't enough to take over the world on its own, so the AI needs to work with humanity in order to effectively pursue its own goals.
Multiple competing AIs form a stable equilibrium keeping each other in check.
One person (or a small group) takes over the world and acts as a benevolent dictator.
At a sufficient level of intelligence, goals converge towards not wanting to harm other creatures/intelligences.
There's a fundamental limit to intelligence that isn't much higher than human level.
Building GAI is impossible because human minds are special somehow.
31
29
14
13
8
3
2
1
OptionVotes
NO
YES
3006
1677
OptionProbability
Other
BDSM
Power couple
"Fun while it lasted"
Me + *Her*
Keynesian beauty contest
Spinning rapidly in opposite directions
Polyamory
Monogamy
Relationship Futarchy
Human + Robot
Enemies to Lovers
Co-conspirators
Philosophically incompatible
Conjoined twinks
Horizontal gene transfer
borgie
Relationship Anarchy
Platonic
Wild Lovers
Friends with Benefits
Relationship constitutional republic: people in it have systemically different levels of power and decision-making influence
Forever alone
Glucose guardian/splenda spender/sugar daddy
Hive / Swarm / Superorganism
Love triangle
Traditional marriage of two cis straight White Christian vanilla people where the man "leads" and is 0-4 years older than the woman, with 2.5 kids (not counting any LGBTQIA+ ones) and a McMansion with 2.5 SUVs/trucks
Universal love
merged consciousness
Rationalussy
Vespertine
liking/reacting to even their most meaningless posts and comments on social media and prediction markets
Matriarchy
Labor Union
Socratic
Partners in crime, like Caroline Ellison and Sam Bankman-Fried
One with an IPO (Intimate Partnership Offer)
Just being one helluva slut
Beard
Paradixical
Friends with Detriments
NP-complete
"I'll never admit to anyone that we met on Tinder"
Hell's Kitchen
Stockholm Syndrome
London Syndrome
Financial Domination
The Price is Right
Crows
Single by choice (except it's other people's choice)
Budget horse
assortative
Hubris quest
Fiends with benefits
Envenomation
Polygamy
Polyandry
Ethical Non-Monogamy
Unethical Non-Monogamy
Celibacy
Aromanticism
Solo poly
Open relationship
Polyfidelity
What's a "relationship"?
Reflexive
Symmetric
Transitive
Metamour
Socially enforced monogamy
Unreciprocated love
Asexual romance
Abusive
"This is my emotional support Ex"
FWB
Enemies-To-Lovers
Codependent
Transference
Monogamish
Voyeuristic
Hobbesian
Romantic Mutual Suicide
Mutuals
Animalistic
Human + AI
parasocial
Polycule
Cuckold
Throuple
Boyce-Codd normal form
Aristotelian
I like big butts and I cannot lie
I like big butts and I cannot tell the truth, how will you escape our dungeon
The Apprentice
15
8
4
3
3
3
3
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
OptionVotes
NO
YES
12522
7986
OptionVotes
NO
YES
1266
814
