OpenAI announces something called "GPT-4.5" or very close to it

'-4.5':6 'announc':2 'call':4 'close':9 'gpt':5 'openai':1 'someth':3

OpenAI announces something new related to Sora

'announc':2 'new':4 'openai':1 'relat':5 'someth':3 'sora':7

OpenAI announces something new related to Voice Engine

'announc':2 'engin':8 'new':4 'openai':1 'relat':5 'someth':3 'voic':7

OpenAI announces something new related to DALL-E

'announc':2 'dall':8 'dall-':7 'e':9 'new':4 'openai':1 'relat':5 'someth':3

'altman':2 'magic/magical/etc':4 'sam':1 'say':3

OpenAI announces something new related to music generation

'announc':2 'generat':8 'music':7 'new':4 'openai':1 'relat':5 'someth':3

Immediate or scheduled public rollout (even if gradual) of Sora

'even':6 'gradual':8 'immedi':1 'public':4 'rollout':5 'schedul':3 'sora':10

'chatbot':6 'gpt2':5 'gpt2-chatbot':4 'mention':2 'someon':1

Update on the partnership with Figure/humanoid robots

'figure/humanoid':6 'partnership':4 'robot':7 'updat':1

OpenAI announces an improvement to GPT-4 similiar to GPT-4-Turbo, which is not major enough to be called GPT-4.5 or GPT-5

'-4':7,11 '-4.5':22 '-5':25 'announc':2 'call':20 'enough':17 'gpt':6,10,21,24 'improv':4 'major':16 'openai':1 'similiar':8 'turbo':12

Temporary technical issues (e.g. mic, presentation, lights, ..., not working)

'e.g':4 'issu':3 'light':7 'mic':5 'present':6 'technic':2 'temporari':1 'work':9

Actual presentation (person talking, video, ...) starts before 10:05 AM

'05':9 '10':8 'actual':1 'person':3 'present':2 'start':6 'talk':4 'video':5

"Regulation" is mentioned, in a legal/policy context

'context':7 'legal/policy':6 'mention':3 'regul':1

A tangible product/device is demonstrated

'demonstr':5 'product/device':3 'tangibl':2

Someone from another (not OAI) company joins on stage

'anoth':3 'compani':6 'join':7 'oai':5 'someon':1 'stage':9

Coalition for Content Provenance and Authenticity Steering Committee mentioned

'authent':6 'coalit':1 'committe':8 'content':3 'mention':9 'proven':4 'steer':7

'mention':3 'partnership':2 'stackoverflow':1

'competitor':3 'googl':1 'mention':4 'search':2

Any mention of the redirect from chat.openai.com to chatgpt.com

'chat.openai.com':7 'chatgpt.com':9 'mention':2 'redirect':5

Mention of AI used by researchers (use cases, how it can help them, etc.)

'ai':3 'case':8 'etc':14 'help':12 'mention':1 'research':6 'use':4,7

The LMSys Chatbot Arena is referenced or shown 

'arena':4 'chatbot':3 'lmsys':2 'referenc':6 'shown':8

At least one specific industry mentioned as userbase or a use case for AI tools

'ai':14 'case':12 'industri':5 'least':2 'mention':6 'one':3 'specif':4 'tool':15 'use':11 'userbas':8

Someone says "patience" or "patient" (not as a noun)

'noun':9 'patienc':3 'patient':5 'say':2 'someon':1

Altman lied about what would be featured at the event in his X posts

'altman':1 'event':10 'featur':7 'lie':2 'post':14 'would':5 'x':13

In the market creator's opinion, the average reaction to the event by journalists is "mixed" or negative

'averag':8 'creator':4 'event':12 'journalist':14 'market':3 'mix':16 'negat':18 'opinion':6 'reaction':9

Any type of new software, AI or not, usable by anyone in the United States willing to pay, is released by OpenAI on the day of the event

'ai':6 'anyon':11 'day':25 'event':28 'new':4 'openai':22 'pay':18 'releas':20 'softwar':5 'state':15 'type':2 'unit':14 'usabl':9 'will':16

In the market creator's opinion, a live demonstration of an OpenAI product is marred by a bug

'bug':18 'creator':4 'demonstr':9 'live':8 'mar':15 'market':3 'openai':12 'opinion':6 'product':13

Any of this words: "Agent", "Agents" or "Agentic" are mentioned

Sam Altman says something along the lines of “what we have now is nothing/laughable compared to what is coming later this year/next year.” 

'along':5 'altman':2 'come':19 'compar':15 'later':20 'line':7 'nothing/laughable':14 'sam':1 'say':3 'someth':4 'year':23 'year/next':22

The start of the stream is delayed at least 5 minutes

'5':10 'delay':7 'least':9 'minut':11 'start':2 'stream':5

GPT-4 becomes cheaper (API price reduced)

'-4':2 'api':5 'becom':3 'cheaper':4 'gpt':1 'price':6 'reduc':7

Presenters give away $10,000 or more in total of free stuff as part of presentation

'000':5 '10':4 'away':3 'free':11 'give':2 'part':14 'present':1,16 'stuff':12 'total':9

The official stream pictures someone wearing a mask (for COVID-19 prevention)

'-19':11 'covid':10 'mask':8 'offici':2 'pictur':4 'prevent':12 'someon':5 'stream':3 'wear':6

An OpenAI employee makes a major gaffe (in the market creator's opinion) that is featured on the front page of the New York Times

'creator':11 'employe':3 'featur':16 'front':19 'gaff':7 'major':6 'make':4 'market':10 'new':23 'openai':2 'opinion':13 'page':20 'time':25 'york':24

In general, resolution will err towards resolving things YES in ambiguous situations. If it seems like something might count according to the spirit of the answer but the letter is unclear, it will probably count. If the event is postponed, this market will be extended until the event occurs.

Prop Bets

🤖 AI

🤖 OpenAI

🖥️ Technology

https://x.com/apples_jimmy/status/1787331421964804324

This market resolves Yes if a new standalone search engine product is 

, whether free or paid. This market is not about the release date of such a product.

"A mere extension to the existing ChatGPT product allowing it internet access" would not count for the purposes of this market. The product should emphasize web search as its primary feature; it should not just be a secondary feature of OpenAI's current ChatGPT / GPT-4 product.

Apologies for ambiguities in wording. I have updated the title for clarity.

See related question on if a search product will be 

https://manifold.markets/Ellis/will-openai-release-a-search-compet-0e9608ec5137

See the following question regarding search capabilities integrated into ChatGPT/GPT-4, not necessarily a standalone product:

https://manifold.markets/Ellis/will-improved-search-capabilities-b?r=RWxsaXM

'announc':4 'chines':1 'gpt':3 'languag':2

'announc':4 'gpt':3 'korean':1 'languag':2

Chat GPT-5 release date or timeframe information given

'-5':3 'chat':1 'date':5 'given':9 'gpt':2 'inform':8 'releas':4 'timefram':7

Something related to DALL-E / Image generation

'dall':5 'dall-':4 'e':6 'generat':8 'imag':7 'relat':2 'someth':1

Someone is announced in a new role / position 

'announc':3 'new':6 'posit':8 'role':7 'someon':1

Announcing a partnership with another company

'announc':1 'anoth':5 'compani':6 'partnership':3

'ai':4 'pornographi':5 'relat':2 'someth':1

Something related to economics / finance / stock trading / money management 

'econom':4 'financ':5 'manag':9 'money':8 'relat':2 'someth':1 'stock':6 'trade':7

'chatbot':5 'detail':1 'gpt2':4 'gpt2-chatbot':3

Partnership with the government of a country

Something Manifold finds disappointing (in a poll created after the announcement)

'announc':11 'creat':8 'disappoint':4 'find':3 'manifold':2 'poll':7 'someth':1

A new physical product in the same niche as the Rabbit R1 or the Humane AI Pin

'ai':16 'human':15 'new':2 'nich':8 'physic':3 'pin':17 'product':4 'r1':12 'rabbit':11

Announcement of model with a new "personality", less refusals, or different RLHF

'announc':1 'differ':11 'less':8 'model':3 'new':6 'person':7 'refus':9 'rlhf':12

Partnership on data collection with other companies

'collect':4 'compani':7 'data':3 'partnership':1

https://twitter.com/sama/status/1788989777452408943

Sam Altman

🤖❤️ AI Alignment

OpenAI is working on a Google Search rival

. Will it release to the public by the end of 2024?

This market resolves Yes for a full version of a web search product, whether it's free or paid. Resolves No if the product is never released, or if it's still in alpha/beta.

Google

IE "make me a 120 minute Star Trek / Star Wars crossover". It should be more or less comparable to a big-budget studio film, although it doesn't have to pass a full Turing Test as long as it's pretty good. The AI doesn't have to be available to the public, as long as it's confirmed to exist.

Scott Alexander's 5 year predictions

Technology

AI Media Creation

In "Moving the Eiffel Tower to ROME", a paper claimed to have identified where the fact "The Eiffel Tower is in France" had been stored within GPT-J-6B, in the sense that you could poke the GPT there and make it believe the Eiffel Tower was in Rome.

"The Eiffel Tower is in France" seems (in my personal judgment) like the sort of fact that early AI pioneers could and did represent within GOFAI systems. GPT-J probably does more with that fact - it can for example answer how to get to the Eiffel Tower from Berlin, believing that the Eiffel Tower is in Rome.  But the paper didn't offer neural transparency into how GPT-J gives directions, we don't know the stored patterns for answering that part - just a neural representation of the brute idea that GOFAI pioneers might've represented with in(Eiffel-Tower, Rome).

This market reflects the probability that, in the personal judgment of Eliezer Yudkowsky, anyone will have uncovered any sort of data, pattern, cognitive representation, within a text transformer / large language model (LLM), whose semantic pattern and nature 

 familiar to AI and cognitive science in 2006 (to pick an arbitrary threshold for "before the rise of deep learning").

Also in 2006, somebody might've represented "the Eiffel Tower is in France" by assigning spatial coordinates to the Eiffel Tower and a regional boundary to France.  Idioms like that appear in eg video games long predating 2006.  Nobody has yet identified emergent environmental-spatial-coordinate representations inside a text transformer model, so far as I know; but even if someone did so before the end of 2026 - as mighty a triumph as that would be - it would not (in the personal judgment of Eliezer Yudkowsky) be an instance of somebody finding a cognitive pattern represented inside a text transformer, which pattern was unknown to cognitive science in 2006.

2006 similarly knew about linear regression, k-nearest-neighbor, principle components analysis, etcetera, even though these patterns were considered "statistical learning" rather than "Good-Old-Fashioned AI".  Identifying an emergent kNN algorithm inside an LLM would again 

constitute "understanding via transparency, within an LLM, some pattern and representation of cognition not known in 2006 or earlier".  Likewise for TD-learning and other biologically inspired algorithms, including those considered the domain of neuroscience (from 2006 or earlier).

GOFAI and kNN and similar technologies did not suffice to, say, invent new funny jokes, or carry on a realistic conversation, or do any sort of intellectual labor.  The 

of this proposition, if relevant, is to assert that by end of 2026 we will not be able to grasp any inkling of the cognition inside of LLMs by which they do much 

than AIs could do in 2006; we will not have decoded any cognitive representations inside of LLMs supporting any

 cognitive capabilities original to the era of deep learning

.  We will only be able to hunt down internal cognition of the sort that lets LLMs do more trivial and old-AI-ish cognitive steps, like localizing the Eiffel Tower to France (or Rome); on their way to completing larger and more impressive tasks, incorporating other cognitive steps; whose representations inside the LLM, even if we have some idea of which weights are involved, have not yet been decoded in a way semantically meaningful to a human.

Technical AI Timelines

AI Alignment 

Mechanistic interpretability 

Resolves YES if GPT-5 is publicly released before Jan 1, 2025; resolves NO otherwise.

Update from March 18: If the next OpenAI flagship model deviates from the GPT-N naming scheme 

 that is going to be released from OpenAI (e.g. they completely changed the trajectory of their naming scheme) then I will count it as "GPT-5" regardless.

However, if the next flagship model is not called GPT-5 but still conforms to the OpenAI GPT-N naming scheme (e.g. GPT-4.5) 

there will eventually be a model called GPT-5

 that will be released (i.e. they are going to continue with the same naming scheme) then I will not count it as "GPT-5".

GPT-5 Speculation

ChatGPT

Eurovision 2024

Eurovision

🍿 Entertainment and Pop Culture

🎨 Culture

🌍 World

Europe

Any performance has to be restarted, for any reason

All four ABBA members performing on stage

'abba':3 'four':2 'member':4 'perform':5 'stage':7

An elected politician currently in office speaks

'current':4 'elect':2 'offic':6 'politician':3 'speak':7

"Palestine" is audibly said by someone with a microphone (crowd screams not enough)

'audibl':3 'crowd':10 'enough':13 'microphon':9 'palestin':1 'said':4 'scream':11 'someon':6

Someone mentions the inclusion of nations geographically outside of Europe

'europ':10 'geograph':7 'inclus':4 'mention':2 'nation':6 'outsid':8 'someon':1

'2nd':6 '2x':3 'place':7 'point':4 'winner':1

'1.5':3 '2nd':7 'place':8 'point':5 'winner':1 'x':4

A participant (i.e. not audience) eats on camera (visibly chewing, holding food, or similar)

'audienc':5 'camera':8 'chew':10 'eat':6 'food':12 'hold':11 'i.e':3 'particip':2 'similar':14 'visibl':9

At least one of the "Big five" make it to top 5

'5':12 'big':6 'five':7 'least':2 'make':8 'one':3 'top':11

Last country jury distributing their points could change position #1 (BEFORE audience votes)

'1':10 'audienc':12 'chang':8 'could':7 'countri':2 'distribut':4 'juri':3 'last':1 'point':6 'posit':9 'vote':13

NO Nordic country's jury gives 12 points to another Nordic song 

'12':7 'anoth':10 'countri':3 'give':6 'juri':5 'nordic':2,11 'point':8 'song':12

Neither Greece nor Cyprus gives other country 12 points

'12':8 'countri':7 'cyprus':4 'give':5 'greec':2 'neither':1 'point':9

All Baltic countries give 12 points to another Baltic country

'12':5 'anoth':8 'baltic':2,9 'countri':3,10 'give':4 'point':6

Please only submit questions concerning the May 11 final, not semifinals or things happening outside of the final as shown on TV.

I reserve the right to resolve markets as N/A for any reason, e.g. if I feel like they're off-topic, poorly worded, et.c.

Sweden

🎷 Music

New York Mets beat Milwaukee Brewers - Mar 29

'29':8 'beat':4 'brewer':6 'mar':7 'met':3 'milwauke':5 'new':1 'york':2

Baltimore Orioles beat Los Angeles Angels - Mar 28

'28':8 'angel':5,6 'baltimor':1 'beat':3 'los':4 'mar':7 'oriol':2

Philadelphia Phillies beat Atlanta Braves - Mar 29

'29':7 'atlanta':4 'beat':3 'brave':5 'mar':6 'philadelphia':1 'philli':2

Chicago White Sox beat Detroit Tigers - Mar 28

'28':8 'beat':4 'chicago':1 'detroit':5 'mar':7 'sox':3 'tiger':6 'white':2

Cincinnati Reds beat Washington Nationals - Mar 28

'28':7 'beat':3 'cincinnati':1 'mar':6 'nation':5 'red':2 'washington':4

Houston Astros beat New York Yankees - Mar 28

'28':8 'astro':2 'beat':3 'houston':1 'mar':7 'new':4 'yanke':6 'york':5

Kansas City Royals beat Minnesota Twins - Mar 28

'28':8 'beat':4 'citi':2 'kansa':1 'mar':7 'minnesota':5 'royal':3 'twin':6

Los Angeles Dodgers beat St. Louis Cardinals - Mar 28

'28':9 'angel':2 'beat':4 'cardin':7 'dodger':3 'los':1 'loui':6 'mar':8 'st':5

Miami Marlins beat Pittsburgh Pirates - Mar 28

'28':7 'beat':3 'mar':6 'marlin':2 'miami':1 'pirat':5 'pittsburgh':4

San Diego Padres beat San Francisco Giants - Mar 28

'28':9 'beat':4 'diego':2 'francisco':6 'giant':7 'mar':8 'padr':3 'san':1,5

Tampa Bay Rays beat Toronto Blue Jays - Mar 28

'28':9 'bay':2 'beat':4 'blue':6 'jay':7 'mar':8 'ray':3 'tampa':1 'toronto':5

'28':7 'beat':3 'chicago':4 'cub':5 'mar':6 'ranger':2 'texa':1

Oakland Athletics beat Cleveland Guardians - Mar 28

'28':7 'athlet':2 'beat':3 'cleveland':4 'guardian':5 'mar':6 'oakland':1

Arizona Diamondbacks beat Colorado Rockies - Mar 28

'28':7 'arizona':1 'beat':3 'colorado':4 'diamondback':2 'mar':6 'rocki':5

Seattle Mariners beat Boston Red Sox - Mar 28

'28':8 'beat':3 'boston':4 'mar':7 'marin':2 'red':5 'seattl':1 'sox':6

Tampa Bay Rays beat Toronto Blue Jays - Mar 29

'29':9 'bay':2 'beat':4 'blue':6 'jay':7 'mar':8 'ray':3 'tampa':1 'toronto':5

Miami Marlins beat Pittsburgh Pirates - Mar 29

'29':7 'beat':3 'mar':6 'marlin':2 'miami':1 'pirat':5 'pittsburgh':4

Houston Astros beat New York Yankees - Mar 29

'29':8 'astro':2 'beat':3 'houston':1 'mar':7 'new':4 'yanke':6 'york':5