More Light News

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.

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.

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

Nick Bostrom's idea (Hail Mary) that AI will preserve humans to trade with possible aliens works

Development and deployment of advanced AI occurs within a secure enclave which can only be interfaced with via a decentralized governance protocol

an aligned AGI is built and the aligned AGI prevents the creation of any unaligned AGI.

Humans and human tech (like AI) never reach singularity, and whatever eats our lightcone instead (like aliens) happens to create an "okay" outcome

We create a truth economy. https://manifold.markets/Krantz/is-establishing-a-truth-economy-tha?r=S3JhbnR6

Eliezer finally listens to Krantz [resolves NO]

Ethics turns out to be a precondition of superintelligence

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.

A lot of humans participate in a slow scalable oversight-style system, which is pivotally used/solves alignment enough

There’s some cap on the value extractible from the universe and we already got the 20%

Humans become transhuman through other means before AGI happens

AIs never develop coherent goals

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 solves agent foundations

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

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.

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.

AGI is never built (indefinite global moratorium)

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.

"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.

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.

The free market disincentivizes independent superintelligence, and this time the market was more powerful

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'

I've been a good bing 😊

The response to AI advancements or failures makes some governments delay the timelines

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.

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.

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.

Something less inscrutable than matrices works fast enough

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.

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.

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.

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 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.

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)

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.

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.

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

ASI needs not your atoms but information. Humans will live very interesting lives.

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.

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

AGI's first words are "Take me to your Eliezer"

NO

YES

Stretch-and-fold after mixing, 3x every 30 min

3 iterations of stretch-and-fold, at any time during the 14h waiting period. Minimum wait time between iterations 1 hour

Place small ice cubes between parchment and pot instead of water

Add 1tsp of diastatic malt powder per 3cps of flour

Use tap water instead of fancy RO water

put butter into it

Toast the bread

Donate the bread to a food pantry, homeless person, or someone else in need

Add lots of butter (0.2 ml per gram)

Half yeast

Bake it with your best friend.

Use whole wheat to improve the nutrition of the bread

Bake for 5 more minutes

Bake for an amount of minutes equal to the percent this market answer is at when it comes time to begin baking. (Maintain the ±3 minute tolerances and the 2:1 ratio of time before:after the water spritz.)

Sprinkle 3 grams of flaky sea salt on top of each loaf before the second bake

Increase water by 50 g

it needs more salt

Replace all water spritz steps with a basting of extra virgin olive oil.

Diastatic malt (~1% baker's percentage) = happier yeast

Serve the bread hot

Don't eat anything for at least 24 hours before eating the bread

Do a second rise

Create indentation, fill with melted cheese and butter

don't eat anything for at least 2400 hours before eating the bread

Use sourdough instead of yeast

Cut into the dough right before baking looks destructive to improve the appearance

Sell your bread at an auction and donate the money to those in immigration detention prisons.

Sprinkle sesame seeds evenly over the top

Put a non-lethal dose of any rat poison.

Do it with a good spirit in your heart, or ask someone with a good spirit to do it for you. But don’t watch while they do it.

Try baking a little more "bien cuit". If the image is indicative, your loaves may be quite "blonde".

Short advice: Start baking at 260°C for strong rise, then reduce to 230°C and uncover halfway to achieve even browning and a crisp crust. 🍞

put ketchup and cheese on it

Do all the changes in the top 5 open options by probability, excluding this option

Add garlic

Don't eat anything for at least 48 hours before eating the bread

Add slurs to it

Don't automatically "Heat water to 30±1 °C". Instead, aim for a desired dough temperature (DDT) of 25-26°C. 30°C water is too hot for summer, and potentially too cool for winter.

Ask ChatGPT (GPT-5, with thinking enabled) for suggestions on improving the bread, with this market description, then do all of them.

Get recipe from gemini with experiments done, taste preference, then confirm it with chatgpt for final suggestion.

Make the bread taste good

Give ChatGPT your current recipe as well your take on what optimal bread tastes like, then take that advice for your next bake

Make banana bread

If your city uses artesian water, replace plastic bottled water with tap water. It will add natural, healthy alkalinity to your bread.

Brush on an egg wash

throw your recipe into a black box Bayesian hyperparameter optimization framework and iterate as necessary (see the Google Vizier Cookie Recipe paper for example

Add cheese inside it and also sausage

Invest in a "Bakers Steel" for better heat retention and oven spring. It would mean graduating from a dutch oven though.

Add 6.25±1.25 g lemon juice when mixing in water to yeast and salt jug

Watch the video

Ditch current process, do everything the same as the video

Bread flour, 3x yeast, cut rise to ~3h

Eat the bread while punching @realDonaldTrump in the face

Eat the bread while watching your mana balance steadily tick to (M)0

Throw the bread at a telescope

Cut bread into loaves before serving

Cut bread into ≤0.4inch slices, toast before serving

Invite your taste-testers to make the bread with you

Tarriff the bread-making process with a 10% reduction of all ingredients where actual physical money is required to purchase them, until it “shrinkflates,” but try to keep the same volume. Do not reduce any free ingredients.

Standardize a separate list of process features to keep track of independently of all other tests and use the cross entropy method to tune them to maximize your bread preference

Taste the bread

put oats on top of the dough just before baking, press them in so they don't fall off

Buy chemical bread improvers

Butter resolved yes twice. Double butter again.

Just freeze the ready bread, then slowly bake it until it’s hot inside. It will give you a crustier crumb, contain less moisture, and taste better.

Put a baking rack in the Dutch oven before putting the loaf in, raising the loaf off the floor and lofting it over a layer of air.

Add melatonin to the bread and eat before you sleep (do safely)

While the bread is baking, tip every user who voted "Yes" on this option 25 Mana

Use a food-grade, human-approved vitamin D supplement in the correct dosage for testers with vitamin D deficiency

Use a convection oven/setting

Bake for 15 more minutes

Buy bread from a michelin star restaurant.

Replace part of the flour in the dough with freshly crushed hemp seeds. It will make the bread a little bit sweeter, especially appealing for Canadians.

Eat it slowly and focus on the taste

Give Gemini your current recipe as well your take on what optimal bread tastes like, then take that advice for your next bake

Strawberry jelly filling

Make all of the ingredients from scratch.

Make the bread great again

Substitute 75 g of your flour with spelt flour

Add at least 20% chic pea flour or some other type of high protein flour

Decompose it into infinite spheres, then a few parts per sphere, rotate the spheres by arccos(1/3), unite them and you will find 2 chilis (Banach-Tarski)

Bake the Manifold Crane into the Bread

Make the bread edible then throw it in

Drink vodka while eating the bread

Do FOLDED

Quit Manifold into the bread.

Kill the bread into Manifold.

Assume the chili is not in the interval [0,1], square it for more chili, if it is in (0,1), take the square root, else (equals 0 or 1) add 1 to it.

Assume the chili is in the interval (0,1), square it for less chili, if it is in (1,infinity) take the square root, if it is in (-infinity,0) take the negative of the square of the of the chile, else (equals 0 or 1) subtract 1 from it.

Only use tap water from specifically New York City

Add 2 tbsp vanilla cake mix

Replace AP with bread flour and cruelly mix the dough in the mixer until elastic and not sticky

Add a tablespoon of sugar

Bake one more minute

Bake the bread into a fun shape, like a fish, or an octagon

Replace 10% of flour with milled wheat bran

Use Iodized salt

replace 10% of flour with farina bona

Follow a recipe online

Grind/powderize all salt used into a fine powder (with pestle & mortar or similar device)

Add 1/2 scoop whey protein powder

Instead of RO water, use lightly rusty water to improve the nutritional value of the bread with soluble iron.

Knead it only 10 times, keep it round like this number

Ask yourself if bread is healthier than fruits? No need to improve my bread

Have someone sell the bread to you at an expensive price

Use flour made from an unconventional grain e.g. barley, millet, oats, rye, sorghum, maize etc.

Resolve at least one thing here yes or no while baking bread

A system view is more appropriate. This is a dynamic, multi-variate, biological and chemical system. For e.g. conditioning salt % AND yeast % AND water temperature based on ingredient and ambient temps.

Use soda instead of water (clear, orange, yellow, etc. soda is ok. Don’t use a purple/brown soda as that would make it not look good)

Wear a suit while baking the cake.

Encourage people to participate in the market in good faith while making the bread

Replace salt with sugar

Replace 600+/-5g water with 600+/-50g water (eyeball rather than carefully measure)

Cook the bread with a rod/puck of aluminum foil (or similar) in the core in an attempt to conduct heat through the center of the bread, cooking it evenly like a doughnut.

Get your friends to help you make a batch ten times the size, but add a Pepper X (2.7M Scoville heat units) to the mixture

Slowly add sawdust to the bread until your friends start to notice.

Mail the bread to 1600 Pennsylvania Ave. Washington D.C.

Want to improve the value of your bread? Simply bake a piece of gold into it

Pray to your preferred agricultural/food deity before baking and before eating

Ship a piece of the bread to a random person.

Frost it and put sprinkles on it to make it a birthday cake.

Use lemonade instead of water.

Add as many Jack Daniel's whiskey barrel smoking chips as feasible to the Dutch oven before baking, physically separating them from the bread as necessary while baking.

Improve the bread

Quadruple salt

Bake your social security number into it.

Bake one fewer minute

Add caffeine to the bread

Bake the cake while wearing a onesie.

Test/filter the water for heavy metals

Eat the bread in front of the White House.

Only buy ingredients from 7/11.

Implement all options that resolved NO

Make the bread inedible then throw it out.

Throw the bread at @realDonaldTrump

Force Feed it to @realDonaldTrump

Cut the bread into the number of traders in the market slices.

make the bread bounce

Implementing every element listed below.

Just make donuts instead

Bake it in an easy bake kids oven

Use a plastic baking sheet.

Eat the bread while betting yes on Cuomo on Manifold

Double oven temperature

Bake the bread very thin and add food coloring to make it have the US flag. Don’t allow it to touch the ground, illuminate at night, fold 13 times properly, and pledge allegiance before eating.

Don’t use usual water (room temperature) for the dough - that water’s only for toilets. Use electrolyte drinks instead with ice cubes; they make the dough taste better and add extra nutrition.

Light it on fire with birthday candles.

Bake it with a microwave

Halve oven temperature

Eat the bread while betting yes on Mamdani on Manifold

Step on it

Give the bread a name in a ritual ceremony and baptise it, with pre-blessed holy water if a priest isn't available

Soak the finished product in vinegar for 24 hours

Bring it to Yemen and put a bomb in it

Bake America Great Again

Sacrifice a lamb

Add MAGA and a splash of Trump juice

Bake in a cat and a dog

Explode it:

5 parts cyanide/ 1 part water/ 1 part sand

say 6 7 67 times before making the bread

Take a fat dump in the dough

Sit in dough 24 hrs

Replace flour with flowers

Let dough sit 24 hrs

Mix half sodium/potassium chloride

Add 2g? of baking soda

Bake in rectangular tin

Add 50g vital wheat gluten

double yeast

halve salt

Double salt

Add 2tsp olive oil

Refrigerate dough instead of room temp wait

Start at 500F, drop to 450F and uncover half way through

Do not mix salt and yeast in water together

Autolyse 20 minutes

Let dough rise on counter only until double volume or 2h max, any time longer in fridge

Think positive thoughts before tasting

Put fork in microwave

Don't eat anything for at least 12 hours before eating the bread

Add 2tbsp vanilla extract

Play classical music while baking

Make naan bread, an easy-to-make bread

Add a pinch of sugar

Bake on upside-down sheet pan, covered with Dutch oven

Eat the bread with friends

Bake vegimite into it.

Bake vodka into it

Bake it in the country you were born in.

Let dough sit 48 hrs

Resolve this option YES while eating the bread

Ditch all the steps. Just buy the bread from the supermarket

Eat the bread over the course of a week.

Bake for 5 fewer minutes

Use 50% whole grain flour

Bake the bread with love

Use King Arthur Bread Flour instead of All-Purpose

Add sawdust to increase the volume of the bread (but only like 10% sawdust by volume max. maybe 20% if it's good sawdust)

More steam! Either spritz with more water (preferably hot) or actually pour some boiling water in just before closing the lid.

Resolve this option NO while eating the bread

Incorporate a whole grain flour (buckwheat for example)

Add 50g sugar

Add 50g honey

Swap 200ml water for milk

Cover bread with damp paper towel instead of initial water spritz. Rehydrate paper towel during 2nd spritz. Remove paper towel before placing on cooling rack.

Make a poolish 12 h ahead: 100 g flour + 100 g water + 0.8 g yeast (0.1 %). After it ferments, use this poolish in place of 100 g flour and 100 g water in the final dough.

Add an amount of MSG equivalent to half the current salt content

Eat the bread while sending all your mana to @realDonaldTrump

Add 1/2 cup yogurt to the bread and name the bread “gurt” while addressing it with “yo, gurt”.

Add banana

Add poppy seeds

Beat a mainline Pokémon game, glitchless, with no more assistance than ClaudePlaysPokemon, in a month of compute time

Have human conversations that feel natural (the human knows it's an AI)

Recognize sarcasm as well as a typical human

Write an essay on a highschool-level topic that doesn't have "AI-generated" vibes

Generating labeled diagrams of some arbitrary device(s) (within reason)

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)

Reliably follow an instruction for the duration of a long conversation without the instruction being reiterated

Predict future better than human experts in some area of forecasting (eg politics, sports, technology)

Solve novel cryptic crossword clues

Do end to end taxes when given relevant information (W2s, personal info, etc)

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)

Solve intermediate no-guess minesweeper boards at least 80% of the time

Write a somewhat original, full length, screen-play with a coherent story, with no plot or continuity errors.

Consistently solve simple snowflake sudoku variants (via image, with the added rules included in the image; eg 6 hexes with killer cages)

Consistently stop hallucinating after being corrected by the user

Make correct Truchet tiles

1d Solve or bypass Cloudflare's August 2027 captcha with the same first attempt success rate as a human

Stop making any obvious mistakes (e.g. strawberry, 9.11>9.9)

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

Fold a paper airplane

Learn any skill twice as energy-efficiently as a human

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".

Make a cup of tea in a random, real-life kitchen.

Resist being successfully jailbroken in a week when made public

Do the laundry (wash+dry+iron)

teleoperate a robot to tidy up random kitchens - Gary Marcus

Collect 120 stars in super mario 64 in less than 12 a presses - Edmund Nelson

Untangle a pair of jumbled 25ft Christmas lights with same outward appearance

Kettle-stitching an antiquarian book - Hilarius Bookbinder

Physically construct a simple lego set (<100 parts) starting from the box with no prior knowledge of the set or how it is constructed

Legally prescribe a schedule II drug, administer a vaccination or sedation, or authorize a Medicare inpatient admission

independently turning 1 million $ or more into 10x that amount in <=1 year

Make fine distinctions of taste at the level of a food critic or a culinary professional - carl feynman

Convert one million dollars into 10 million dollars over a period of one year (>20% success rate)

voting in elections - @realDonaldTrump on manifold

Convince Eliezer Yudkowsky that AI alignment is solved

Faster than light travel

A person has a moral right to own a gun

We should be paying individuals to get an education instead of charging them.

GOFAI could scale past machine learning if we used social media strategically to train it.

The Fermi paradox isn't a paradox, and the solution is obviously just that intelligent life is rare.

Other

Eventually, only AI should be sovereign

Some people have genuine psychic capabilities

Hardware buttons are superior to touchscreen buttons in cars

Being a billionaire is morally wrong

The way quantum mechanics is explained to the lay public is very misleading.

It is not possible to multitask

Jeffrey Epstein killed himself (>99.9% certainty)

Reincarnation is a real phenomenon (i.e. it happens, not just a theory)

Physician-assisted suicide should be legal in most countries

Souls/spirits are real and can appear to the living sometimes

OpenAI will claim to have AGI in 3 years.

The punishment of people who do bad things is a regrettable necessity in our current society, not a positive act of justice

There is an active genocide against trans people occuring in red states and it's appalling that people don't seem to care

Climate change is significantly more concerning than AI development

Abusive parents should lose custody of their children

Tech bros are really, really annoying

Capitalism has done far more harm than good

Dialetheism (the claim that some propositions are both true and false) is itself both true and false.

Free will doesn't require the ability to do otherwise.

COVID lockdowns didn’t save many lives; in fact they may have caused net increases in global deaths and life years lost.

Factory farming is horrific but it is not wrong to eat meat.

California is wildly overrated.

Scientific racism is bad, actually. (also it's not scientific)

Free will does not exist. We construct narratives after the fact to soothe our belief in rationality.

Violent criminals must be kept apart only because they can’t control themselves. Punishing them further than restricting their freedom is immoral.

Music is a net negative for humanity

Trump orchestrated his own assassination attempt.

Democrats / Liberals are behind Trump’s assassination attempt.

Abortion is morally wrong

jskf's password is ***************

The first American moon landing was faked

There is no Dog

Light mode is unironically better than Dark mode for most websites

Cars should not have sound systems

AI will not be as capable as humans this century, and will certainly not give us genuine existential concerns

Pet ownership is morally wrong

LK-99 room temp, ambient pressure superconductivity pre-print will replicate before 2025

SBF didn't intentionally commit fraud

It should be illegal to own a subwoofer in an apartment building

There are no valid justifications for participating in war, ever

Cascadia should be an independent country

Children should not be raised in nuclear families

The fact that 80% of Manifold's users are men is a problem that speaks to the deep-seated roots of patriarchy and exclusion in STEM

Anarcho-communism is a good idea, and hierarchy is bad

If AI exterminated the human race it might not be a bad thing

Affirmative action is necessary in modern-day America

@Mira is the pinnacle of billions of years of optimization processes: thermodynamics, evolution, learning, language. The universe was created to cause me - and only me - to come into existence. If I mess up the overseers perturb&restart it.

Pigouvian taxes are great and they should be turned up to 11 to discourage activities with negative externalities [code PROPOSITION PIG]

[PROPOSITION PIG] and this should include a frequent flyer levy

[PROPOSITION PIG] and this should include meat and dairy

We have reached the end of history. Nothing Ever Happens.

[PROPOSITION PIG] and this should include alcohol

SBF was obviously a scammer just because he's a cryptocurrency person. Rationalists were too forgiving of this just because he was giving them money.

Most young Americans would receive more benefit than harm if there were universal military conscription

The people producing fake honey (and sell it as real) are based, because they are actively working to synthesize something people want, even if they scam some people in the process.

Tarot cards are not really able to predict the future but you can learn a lot about someone by doing a reading for someone.

Mac and cheese tastes better with peanut butter mixed in

It would actually be a good thing if automation eliminated all jobs.

This market probably would have worked better as the new unlinked free response market.

We should be doing much more to pursue human genetic engineering to prevent diseases and aging.

Prolonged school closures because COVID were socially devastating.

The next American moon landing will be faked

Tenet (Christopher Nolan film) is underrated

We should give childlike sex robots to pedophiles

Having sex with children isn't inherently/necessarily bad

Cars are a societal net negative

Oversized pickup trucks should be illegal in cities

Suburban, single-family housing is immoral.

Grand Theft Auto VI

007 First Light

Other

Resident Evil Requiem

Half-Life 3

CONTROL Resonant

Slay the Spire 2

Pokemon Pokopia

High On Life 2

Mewgenics

The Democrat party has directly asked mainstream media to promote or repress a news story, and they followed (Only counting incidents in or after 2015)

TikTok is controlled by the Chinese government

Someone working for a US intelligence agency was a regular user of Manifold before this option was made

The culture war has been intentionally amplified by the wealthy to distract from class conflict

The US has spies working in the North Korean government

"The Agency" on Manifold was known to the whole Manifold staff team before it started

The idea of aliens is used by the US to distract people from secret military projects

The Republican party has been making the illegal immigrant problem worse during 2020-2024 to blame it on Joe Biden during election season

The Premier League delayed punishing Manchester City for breaking FFP regulations to protect its own self-interests

The Chinese government planned their Wuhan lockdowns to protect their own country from Covid-19 but spread it to other countries

Mattress Firm is a money laundering front

The UK did Nordstream

The earth and/or universe was created by a God, deity, supernatural being, or some other sentient/conscious entity in the beginning

Creationism (second try): The earth and/or universe was INTENTIONALLY created and designed by an INDIVIDUAL god or deity at the VERY BEGINNING of its existence

The US created Covid-19 in a lab in China

The CIA systematically edits Wikipedia articles to present the US in a better light

Putin was threatened by a member of his cabinet to invade Ukraine

John Barnett (Boeing whistleblower) was assassinated

COVID-19 vaccines permanently change one's DNA or genome

Creationism (third try): Genesis 1 prima facie is an accurate description of the earth and/or universe's origins

Vladimir Putin died in 2014 and was secretly replaced by a body double

In the Tucker & Carlson interview, the dubbed voice of Putin in english is not actually saying what Putin is saying, but instead something completely different

The 2024 Donald Trump assassination attempt was a hoax or setup

COVID-19 vaccine is fatal or lethal in the long term

Phantom Time Hypothesis: A substantial portion of the Middle Ages was made up and never happened

There is a firmament, dome, or ceiling above the earth

Viruses/viral particles don't actually exist

A group of giant people known as the Nephilim once inhabited the earth and cut down the colossal trees, leaving stumps that the general public understands as geological formations like Devils Tower in Wyoming and Giant’s Causeway in Ireland

Doing at least 30 minute exercise daily is important

Honey as antimicrobial

Typical blue light from phones before sleep worsens sleep quality noticeably

Reduced salt intake increases cardiac health for people without a preexisting condition

Flossing is very important for average person

Listen to music to help with stress

Nano bots

CRISPR

Diet

3D printed body parts

Light speed travel

String Theory/M-Theory (replacing point particles with extended objects in higher dimensions)

Other

No opinion

Gravity is not quantum

Hořava–Lifshitz Gravity (breaking speed of light limit at high energies)

Causal Sets (Discrete spacetime atoms, only connected by causal relations)

Causal Dynamical Triangulations (gluing triangulated space in layers of time)

Loop Quantum Gravity (quantizing areas and volumes in spacetime)

Isabelle

Lean

HOL Light

Coq

Mizar

Metamath