Goalposts

Houshalter, 2016-01-19, about whether video games or language understanding are the path to AGI

mcv, 2016-01-19, about whether game AI can master strategy and incomplete information, beyond fast reflexes

kang, 2016-02-07, about a quiz telling computer-generated paintings apart from human artists' work

adrusi, 2016-02-21, about whether future AI could detect and strip all advertising, ending the ad-blocking wars

Houshalter, 2016-03-09, about whether the Turing test's goalposts have moved as AlphaGo beats Lee Sedol

TillE, 2016-03-10, about whether AI will write popular books or make whole animated films, after AlphaGo's win

kerkeslager, 2016-03-11, about whether AlphaGo beating Lee Sedol is a step toward general-purpose AI

daveguy, 2016-03-12, about whether AI could actually pass the Turing test after AlphaGo's wins

vamur, 2016-03-12, about which game AI should tackle next after AlphaGo beat Lee Sedol

SilasX, 2016-03-14, about whether AI can attempt arbitrary problems from a plain-language spec

kleer001, 2016-03-17, about what the next concrete AI milestone should be after AlphaGo beat Go

thangalin, 2016-03-27, about whether any AI can answer commonsense questions like which object shattered

undersuit, 2016-03-29, about sponsored content in journalism and whether AI could detect hidden ads

tremon, 2016-03-31, about whether word statistics alone can solve Winograd schema pronoun puzzles

fizzbatter, 2016-05-18, about whether AI can turn plain English into structured, actionable data and back

x1798DE, 2016-06-10, about whether AI can understand books well enough to answer detailed questions

calgoo, 2016-06-22, about an article claiming coding is over and apps can be made without programming

lottin, 2016-08-01, about whether AI is less impressive than claimed, after an AI beat a professional Go player

Grue3, 2016-08-01, about the claim that AI is always less impressive than claimed

Houshalter, 2016-08-10, about whether AI can become intelligent from text alone, without senses like sight

seanwilson, 2016-08-19, about the AI effect, where solved problems stop counting as real intelligence

MawNicker, 2016-08-22, about whether a neural network passing the Turing test would reveal how humans think

YeGoblynQueenne, 2016-09-18, about whether Google's self-driving car record proves it drives better than humans

Houshalter, 2016-09-22, about whether "seed AI" that builds stronger AIs could lead to general AI

MrQuincle, 2016-09-23, about whether general object recognition was ever considered the holy grail of AI

M_Grey, 2016-09-29, about what AI would need to show before deserving rights

crispyambulance, 2016-10-26, about AI style transfer images and whether their low resolution loses painterly detail

ainot, 2016-11-05, about whether AGI will eventually take over software design work like UML modeling

sweezyjeezy, 2016-11-07, about a machine learning rap lyric generator and whether it disrespects hip hop

thenomad, 2016-11-26, about whether games against bots could offer real PvP thrills without Dota's toxic players

eva1984, 2017-01-02, about whether AI hype in 2016 was overblown compared to real NLP progress

jordigh, 2017-01-04, about a TensorFlow project that drives Mario Kart on an easy course

Hydraulix989, 2017-01-04, about whether AI and machine learning will revolutionize the legal profession

sethbannon, 2017-01-31, about whether a poker AI beating top pros counts when engineers tuned it nightly

fauigerzigerk, 2017-02-15, about whether AI could build apps from spoken wishes like "make a web page"

Houshalter, 2017-02-26, about how to define high-level machine intelligence and AGI

Houshalter, 2017-02-28, about how to define high-level machine intelligence concretely

nradov, 2017-03-30, about whether general AI will leave humans with no more problems to solve or jobs

joeyspn, 2017-04-03, about why chatbots have struggled to live up to the hype

Insanity, 2017-04-04, about whether shooter game bots can feel like real human players

dreamcompiler, 2017-04-11, about whether neural network decisions need to be explainable in human terms

wikwocket, 2017-04-28, about whether junior developers will be replaced by deep learning within 10-15 years

madamelic, 2017-05-03, about whether an email assistant that learns which emails you answer counts as AI

nathan_f77, 2017-05-27, about whether AlphaGo's Go wins signal intelligent machines will upend all intellectual work

codeulike, 2017-05-28, about whether intelligent game characters deserve any moral consideration

d--b, 2017-07-07, about whether chatbots can deliver good customer service

jon_richards, 2017-07-30, about whether marketing will remain a human job in a robot economy

flunhat, 2017-08-12, about whether OpenAI's Dota 2 bot shows AI goalposts are being moved

flunhat, 2017-08-12, about whether critics keep moving AI goalposts after OpenAI's Dota 2 bot

cthor, 2017-08-13, about whether OpenAI's 1v1 Dota 2 bot is real progress or hype

mabramo, 2017-08-17, about OpenAI's 1v1 Dota 2 bot and whether it is really impressive

moxious, 2017-08-23, about whether computers are still dumb or AI goalposts keep moving

huhlig, 2017-09-03, about whether customer-service chatbots can hold natural, human-like conversations

komaromy, 2017-09-17, about which next big advance in AI people are most looking forward to

moxious, 2017-09-29, about whether the Turing test is a meaningful test of machine intelligence

paganel, 2017-10-06, about whether superintelligent AI would mostly face mundane real-world problems rather than breakthroughs

Synaesthesia, 2017-10-14, about whether AGI can be defined and how far AI is from it

alistproducer2, 2017-10-16, about a GM executive calling Musk's claim of Level 5 driving without lidar impossible

yters, 2017-11-06, about a report on efforts to build AI systems that can build AI

modi15, 2017-11-11, about whether new neural network research brings dangerous, out-of-control AI close

revel, 2017-11-13, about how close computers are to true conversational ability

JoshTriplett, 2017-11-19, about wanting a voice-interactive AI dungeon master for playing D&D

jxdxbx, 2018-01-29, about what a truly useful AI voice assistant should be able to do

klibertp, 2018-02-10, about whether AI will soon take away programming jobs

Grue3, 2018-05-10, about whether AI that can trick humans, like Google Duplex, must disclose itself

romaniv, 2018-05-15, about which game OpenAI should tackle after Dota, such as CS:GO

whataretensors, 2018-05-19, about whether Google Duplex should tell people they're talking to a machine

apk17, 2018-05-19, about whether passing the Turing test in our lifetime is exciting, after Google Duplex

tim333, 2018-05-21, about whether Google Duplex is a notable advance or just overhyped tech news

Omnus, 2018-06-06, about whether an algorithmic music project really counts as AI writing its own music

AstralStorm, 2018-06-25, about OpenAI's Dota 2 bots trained on 180 years of games per day

dselsam, 2018-07-03, about whether AI beating humans at chess, Go and proofs means intelligence is easier than gripping

pacavaca, 2018-07-08, about tech companies quietly using humans to do work sold as AI

hadsed, 2018-07-30, about whether AI could learn how software systems work from their behavior

chongli, 2018-08-21, about whether Facebook and Google can judge truth and trustworthiness automatically instead of relying on authority

abdullahkhalids, 2018-09-12, about criticism of OpenAI Five's Dota 2 bots and whether they play the real game

TillE, 2018-09-18, about iOS 12 mistranslating "I missed my flight" into Japanese

thaniri, 2018-09-21, about Tencent's AI winning over 90 percent of StarCraft II games

Simulacra, 2018-10-12, about whether autonomous cars can ever cope with unpredictable human drivers

nradov, 2018-10-15, about whether simulating a mapped connectome could produce human-level general intelligence

esotericn, 2018-10-29, about manipulated YouTube comments and fears that bots online threaten democracy

jobigoud, 2018-11-12, about whether CAPTCHAs, as reverse Turing tests, show strong AI hasn't arrived yet

heyjudy, 2018-12-31, about whether AI could do developers' and designers' whole job end to end

chubot, 2019-01-07, about Waymo's CEO saying autonomous cars will never drive in all conditions

orangeeater, 2019-01-19, about a 1984 rebuttal to Searle's Chinese Room argument about machine understanding

tim333, 2019-01-22, about whether Kurzweil's predictions that AI will surpass human intelligence are falsifiable

rcthompson, 2019-01-25, about whether AlphaStar will soon beat humans at StarCraft II under any fair handicap

sethammons, 2019-01-30, about whether simple-minded developers will soon be replaced by moderately intelligent AI

eschneider, 2019-02-13, about people ignoring real calls, like a hospital's, because they assume they're robocalls

Retra, 2019-02-15, about whether GPT-2 tuned on movie subtitle dialogue could pass a proper Turing test

subroutine, 2019-03-14, about whether AI is anywhere near human-level, judged by Siri's limits

TicklishTiger, 2019-03-27, about whether the Turing award or the Turing test matters more, after deep learning pioneers win

londons_explore, 2019-04-09, about whether AI could fly a crippled aircraft better than human pilots

roywiggins, 2019-06-15, about a paper arguing that Turing machines can never pass the Turing test

sneak, 2019-07-05, about California's law banning bots from posing as real people without disclosure

sixplusone, 2019-07-17, about Waymo's CEO saying fully autonomous level five cars will basically never exist

ilaksh, 2019-07-20, about an Amazon warehouse worker's account and whether robots could replace them

sgt101, 2019-07-24, about whether passing the Turing test would show that a machine is conscious

smt88, 2019-08-23, about whether forums like HN could quietly fill up with polite, human-seeming bots

CPLX, 2019-09-29, about collisions with Tesla's Smart Summon and whether more data leads to self-driving

pjc50, 2019-10-11, about whether people can form real emotional connections with AIs rather than humans

eternalny1, 2019-10-26, about how soon AI that beats humans at programming contests would replace programmers

taneq, 2019-11-05, about whether GPT-2 successors will soon write essays good enough for students to cheat

School-Cotton, 2019-11-07, about whether riding a bike is harder for a robot than driving a car

username90, 2019-11-14, about what would count as artificial general intelligence, beyond passing the Turing test

diek, 2020-01-30, about why someone chose a Chevy Bolt over a self-driving-capable Tesla

winternett, 2020-02-03, about Tesla's self-driving overlay video and whether rule-based driving systems count as AI

fnord77, 2020-02-15, about self-driving cars turning out to be harder than expected

dmitrygr, 2020-02-20, about Tesla Autopilot being tricked into speeding by a modified speed sign

leftyted, 2020-02-27, about whether the Turing test is a good measure of human-level AI

JohnJamesRambo, 2020-03-04, about whether AI is anywhere near ruling humans, given it can't drive in snow

djsumdog, 2020-05-14, about whether self-driving cars will ever be viable, after Cruise layoffs

ben_w, 2020-05-20, about an OpenAI code-generation demo and whether programmers will lose their jobs

pjc50, 2020-05-22, about Siri's poor answers to questions like what time it is in London

mixedbit, 2020-06-02, about board games with simple rules but complex gameplay, like Dixit

memexy, 2020-06-04, about whether AI is anywhere near human thinking, and whether to fear it

Barrin92, 2020-06-11, about whether the OpenAI API's text models mean AGI is only a few years away

rvz, 2020-07-19, about whether GPT-3's surprising results mean AGI may be closer than skeptics think

vsskanth, 2020-07-20, about why GPT-3 is a big deal and what GPT-4 might do

MayeulC, 2020-07-20, about what GPT-3 might do next, after it generated a presentation from a prompt

flukus, 2020-07-27, about whether autonomous trains and planes can handle random obstacles in open systems

rvz, 2020-08-15, about whether GPT-3 could help humanity fight Covid-19

thisisbrians, 2020-08-20, about whether any new kind of work will be immune to automation, in a universal basic income thread

isoprophlex, 2020-09-10, about a claim of Latin in the Voynich manuscript and whether GPT-3 could translate it

username90, 2020-10-25, about what intelligence tests would be equally valid for humans, animals and machines

stanrivers, 2020-11-02, about a South Park satire series using AI deepfakes and what deepfakes mean for trusting video

PaulHoule, 2020-12-16, about whether lacking copyrighted novels in training data limits AI-generated fiction and personas

jamil7, 2020-12-20, about whether cooking is easy to automate, as argued in a thread on abolishing work

simonh, 2020-12-30, about whether the hard part of the Turing test is intelligence or imitating human quirks

mhh__, 2021-01-03, about an AI-powered landing page builder and whether AI generation counts as design

jboggan, 2021-03-05, about an AI that learned no-communication "gunboat" Diplomacy rather than the full game

mschuster91, 2021-04-15, about whether AI can replace the human content moderators Facebook relies on

mullingitover, 2021-05-06, about AI tutoring as a way to give every student one-on-one teaching

PoignardAzur, 2021-06-09, about the release of the open GPT-J-6B language model

KMag, 2021-06-15, about whether Google's AI floorplanning tool means AI now designs chips

username90, 2021-06-24, about AI code generation with the GPT-J-6B language model

qqqwerty, 2021-06-29, about whether GitHub Copilot's code suggestions will meaningfully change programming work

rowanG077, 2021-07-02, about whether AI can write large software systems, given GitHub Copilot and AI-written text

computerex, 2021-07-04, about whether Copilot's text-based approach can lead to general program synthesis

Houshalter, 2021-07-06, about whether GitHub Copilot is really intelligent or just a code search engine

h0l0cube, 2021-08-04, about whether game development can be automated by AI

jonesn11, 2021-08-12, about how well OpenAI Codex handles real programming beyond its demos

drooby, 2021-09-15, about whether self-driving on unmapped local roads requires general AI

jwommack, 2021-09-22, about an article saying AI writes code like humans, bugs included

hutzlibu, 2021-11-25, about what early signs would indicate an approaching AI singularity

hutzlibu, 2021-11-26, about early signs of a singularity and whether the Turing test still matters

bryan0, 2021-11-27, about early signs of a singularity and whether the Turing test was already passed

genewitch, 2021-11-29, about adversarial image attacks and whether self-driving cars are really coming soon

nickpp, 2021-12-02, about why search results are flooded with SEO spam and auto-generated pages

johnisgood, 2021-12-05, about a suicide capsule maker's plan to use AI instead of psychiatrists to assess mental capacity

bryan0, 2021-12-16, about whether the Turing test still means anything as a test of AI understanding

Accujack, 2022-01-06, about a US law requiring new cars to detect impaired drivers and limit driving

majormajor, 2022-01-06, about whether 'AI' is a moving target as chatbots keep improving

AstralStorm, 2022-01-14, about whether an AI that reproduced a known theorem-solving algorithm counts as science

37ef_ced3, 2022-01-19, about whether coding challenges could work as CAPTCHAs that Copilot can't solve

bovermyer, 2022-01-20, about whether AI dubbing could replace foreign-language subtitles in films

Jasper_, 2022-01-24, about what neural-network AGI would need, and whether self-driving cars can handle social roads

diehunde, 2022-02-02, about whether AlphaCode and Copilot will make programming obsolete

qualudeheart, 2022-02-02, about whether AlphaCode threatens software engineers whose job is solving business problems

bricemo, 2022-02-02, about DeepMind's AlphaCode and whether AI reaches top-1% competitive programming within five years

ahgamut, 2022-02-02, about whether AlphaCode can solve kinds of programming problems absent from its training data

ensan, 2022-02-02, about DeepMind's AlphaCode solving competitive programming problems

jowday, 2022-02-28, about whether camera-only self-driving cars need human-like scene understanding beyond image recognition

type_enthusiast, 2022-03-04, about a $10,000 bet on whether fully self-driving cars will exist by 2030

Beldin, 2022-03-05, about what AI driving must achieve for fully self-driving cars by 2030

jll29, 2022-03-19, about DARPA's view of AI and how far it is from real intelligence

virgilp, 2022-04-01, about whether AI goalposts have moved since beating Go was thought to require general intelligence

Rochus, 2022-04-05, about what would make AI-generated music performance convincing to musicians

chrisco255, 2022-04-06, about whether AI like DALL-E 2 signals that CRUD programming will be automated

6gvONxR4sf7o, 2022-05-07, about whether we will ever agree that generalized AI exists, given that hard problems get solved by clever search

tuatoru, 2022-05-11, about whether robots can handle the physical skill involved in construction work

px43, 2022-05-12, about what bar an AI would need to clear to count as AGI, after DeepMind's Gato

Jensson, 2022-05-12, about whether a generalist AI like DeepMind's Gato could apply for and hold a job

photochemsyn, 2022-05-20, about which games AI hasn't mastered, and whether it could pass as a D&D player

samtimalsina, 2022-05-24, about a story about AI written by one AI and illustrated by another

nautilus12, 2022-05-24, about whether humor and irony are what GPT-3 still can't do

mark_l_watson, 2022-06-07, about whether embodiment is necessary for AGI, or whether interacting with people is enough

xaduha, 2022-06-12, about whether Tesla's camera-only self-driving lags behind lidar-based robotaxi companies

_olhy, 2022-06-12, about whether LaMDA seeming alive means it is sentient, after a Google engineer's claim

tsimionescu, 2022-06-13, about whether LaMDA could be sentient if it fooled a skeptic in a Turing-style test

ecshafer, 2022-06-13, about whether sentience is a neuroscience question and how to test chatbots like LaMDA for it

mkaic, 2022-06-13, about whether we know enough about sentience to judge if AI like LaMDA is sentient

Group_B, 2022-06-13, about whether chatbots are now good enough that people need worry less about dying alone

nradov, 2022-06-18, about whether language models fooling some people means they are close to passing the Turing test

dane-pgp, 2022-06-19, about whether Google's LaMDA really passed the Turing test, as an article claimed

the_d3f4ult, 2022-06-22, about why AI hasn't replaced highly paid brain surgeons as Alphabet pushes into healthcare

paganel, 2022-06-27, about whether AI-generated fake identities will become indistinguishable from real people online

zaptrem, 2022-07-13, about whether Google's LaMDA is sentient and how people judge sentience in others

yrgulation, 2022-07-24, about a prediction of when software development jobs will be automated

Traubenfuchs, 2022-08-07, about whether progress in AI and self-driving has stalled, under a story on Uber's losses

hef19898, 2022-08-17, about whether AI-generated art, films and novels will replace human artists

ianceicys, 2022-08-19, about John Carmack's new AGI company Keen Technologies raising $20M

ryanSrich, 2022-08-19, about John Carmack's AGI company and whether an AGI could break out of its simulation

nradov, 2022-08-20, about how to define AGI concretely, prompted by John Carmack's new AGI company

archagon, 2022-08-22, about whether image generators like Stable Diffusion threaten studios like Marvel

dangerwill, 2022-08-25, about how AI-generated fakery and content threaten creative and programming jobs

roywiggins, 2022-08-30, about whether image AIs like DALL-E create new things the way human artists do

Retric, 2022-08-30, about whether automation leaves room for a long tail of human creators

roywiggins, 2022-08-30, about whether DALL-E can innovate the way artists like van Gogh did

imglorp, 2022-10-01, about whether AI will replace software developers

thwayunion, 2022-10-12, about whether language models should be expected to do math at all

nuker, 2022-10-16, about whether self-driving cars need reasoning that ML systems lack

oldstrangers, 2022-10-21, about how AI text generators will change education and student learning

ec109685, 2022-10-31, about whether humans are good drivers, in a thread on Tesla dropping radar and ultrasonic sensors

jakeinspace, 2022-11-01, about how to measure progress toward artificial general intelligence with benchmarks

giancarlostoro, 2022-11-01, about the AI effect and what people would actually count as AI

dinkblam, 2022-11-13, about whether language models like GPT-3 are nearly AGI and critics keep shifting the bar

bscphil, 2022-11-17, about research claiming to reconstruct images people see from non-invasive brain scans

darepublic, 2022-12-04, about whether AI placing first in Advent of Code takes the fun out of programming

dvt, 2022-12-04, about whether ChatGPT is intelligent, prompted by a git repo written entirely by it

Arch-TK, 2022-12-04, about whether ChatGPT can write Linux kernel device drivers

sumtechguy, 2022-12-05, about whether AI-assisted coding is the start of the end of software development jobs

inopinatus, 2022-12-06, about whether nitpicking ChatGPT's failings misses how fast AI is progressing

dylan604, 2022-12-06, about AI voice modification making call center workers sound American

gsail11, 2022-12-07, about how ChatGPT will affect software careers and junior developers

constantcrying, 2022-12-11, about whether Copilot-style tools threaten programmers or are just like Stack Overflow

contravariant, 2022-12-12, about whether HN should ban comments generated by ChatGPT

GoblinSlayer, 2022-12-15, about whether AI-generated code that is 95% right is good enough, amid AI art fights

runarberg, 2022-12-17, about whether ChatGPT's attempts to explain jokes show it understands humor

borbulon, 2022-12-19, about a claim that AI will soon take over software engineering

dublin, 2022-12-20, about configuring firewalls such as pfSense and OPNsense

mannykannot, 2022-12-25, about whether the Turing test is a good criterion for machine understanding

remarkEon, 2022-12-25, about whether AI answer engines summarizing news will undercut news outlets and original reporting

denton-scratch, 2023-01-03, about an AI tool that summarizes legal documents such as Google's terms

lolinder, 2023-01-04, about whether ChatGPT in Bing would be useful if it could say "I don't know"

zone411, 2023-01-05, about which problems would be easy for a superhuman AGI but very hard for humans

pavlov, 2023-01-07, about Gary Marcus's skepticism that ChatGPT understands, e.g. a robot tidying a room

purpleblue, 2023-01-12, about whether a GPT-3-like model trained on astronomy research could explain anomalies

_Algernon_, 2023-01-12, about whether a GPT-3-like model trained on astronomy research could answer open cosmology questions

random_upvoter, 2023-01-13, about whether ChatGPT-written essays harm education and how deep its knowledge really runs

zactato, 2023-01-20, about a tool that uses GPT-3 to write git commit messages automatically

dragonwriter, 2023-01-20, about a tool that uses GPT-3 to write git commit messages automatically

zephrx1111, 2023-01-20, about the worry that an AI might fail the Turing test on purpose

squidbot, 2023-01-22, about whether AI program generation like ChatGPT threatens software developers' jobs

nostromo123, 2023-01-22, about what ChatGPT and AI code generation mean for software engineering jobs

iLoveOncall, 2023-01-22, about whether ChatGPT can challenge Google search despite often being wrong

kieselguhr_kid, 2023-01-25, about whether GPT prompts can replace SQL analysts and software developers

AnimalMuppet, 2023-01-25, about an article claiming human-level AI translation by 2030 would mean the singularity

bcbrown, 2023-01-26, about whether GPT-written tests passing on the first try shows they are good tests

dinkumthinkum, 2023-01-26, about whether ChatGPT's answers to silly prompts mean it could handle real legal cases

tremon, 2023-01-26, about Marines beating a DARPA robot by hiding under a box, and AI lacking common sense

hbrn, 2023-01-26, about whether ChatGPT has no thought of its own and only predicts likely responses

gamblor956, 2023-01-26, about an AI lawyer that was blocked from arguing in court

andrewstuart, 2023-01-27, about whether ChatGPT can summarize a big, unfamiliar codebase

josephg, 2023-01-29, about what benchmark for human-level AI people would still accept once it is beaten

logarhythmic, 2023-01-29, about whether AI will soon overcome the difficulties of software engineering

airbreather, 2023-01-31, about what programming languages are for once AI can write all code

HarHarVeryFunny, 2023-02-03, about whether ChatGPT really knows things or just makes them up

bjackman, 2023-02-04, about DocsGPT, an open-source AI assistant that answers questions from library documentation

pixl97, 2023-02-04, about whether headlines about ChatGPT passing tests and interviews mean anything

jhoelzel, 2023-02-07, about whether AI coding assistants like Copilot work beyond simple, context-free functions

hot_gril, 2023-02-07, about whether AI can be funny, after an AI Seinfeld show was banned for its jokes

userbinator, 2023-02-07, about Microsoft launching a new Bing search powered by an upgraded ChatGPT

quaestio, 2023-02-09, about AI fatigue and whether ChatGPT is reliable enough for engineering work

pegasus, 2023-02-10, about whether large language models show theory of mind, and how to test one-shot learning

hawkesnest, 2023-02-10, about whether ChatGPT could replace human programming mentors

gfodor, 2023-02-13, about a Clojure DSL for full-stack web UIs and whether it could help AI write apps

WoodenChair, 2023-02-15, about whether humans are just more advanced next-word predictors, as Bing chat is said to be

imiric, 2023-02-19, about whether GPT-style AI makes design tools and web pages outdated

mcv, 2023-02-19, about what it would take for an AI to convince people it is conscious

RC_ITR, 2023-02-21, about whether GPT multiplying 4-digit numbers would show it learned arithmetic

knaik94, 2023-02-26, about how realistic AI-generated audio like AudioLM can become

rnosov, 2023-02-28, about what evidence would convince someone that an AI is sentient

resource0x, 2023-03-07, about whether the Turing test is still the real goalpost for AGI

ghaff, 2023-03-09, about computer vision in manufacturing versus robots that do household chores

nopinsight, 2023-03-14, about GPT-4 acing standardized exams but struggling with competitive programming and AMC

audunw, 2023-03-14, about whether GPT-4's better test scores mean fewer confidently wrong answers

nradov, 2023-03-14, about whether people are moving the goalposts on AGI after GPT-4's release

Jevon23, 2023-03-14, about whether people are moving the goalposts on AGI as GPT-4 arrives

camjohnson26, 2023-03-14, about whether GPT-4 getting good at math should worry people more than its exam scores do

TaupeRanger, 2023-03-14, about whether skeptics keep moving the goalposts as GPT-4 improves

JTyQZSnP3cQGa8B, 2023-03-14, about whether GPT-4 writing code from pasted specs makes software engineers unnecessary

RC_ITR, 2023-03-14, about whether Othello-GPT shows that LLMs learn world models rather than just statistics

nine_k, 2023-03-15, about an AI posing as a human online, after GPT-4 hired a TaskRabbit worker

stefl14, 2023-03-15, about whether transformers being Turing complete means LLMs can learn real algorithms

nradov, 2023-03-15, about how to define AGI when nobody agrees on what the term means

wizofaus, 2023-03-15, about whether LLMs like GPT-4 are general enough to count as AGI

yosito, 2023-03-17, about testing GPT-4's code-writing ability on real-world problems

qlm, 2023-03-17, about whether language models truly understand things or only generate plausible text

qlm, 2023-03-17, about whether LLMs truly understand things or can invent things the way humans do

marcosdumay, 2023-03-19, about whether AI will reduce how many software developers are needed

sarchertech, 2023-03-20, about whether AI will automate founders' work as well as software engineers'

HeavyFeather, 2023-03-22, about ChatGPT submitting a pull request to Rails and whether that is ordinary

tylerhou, 2023-03-22, about whether AI models could be used to invert hash functions

bsder, 2023-03-23, about whether LLMs could make programming interesting again by handling boilerplate

raydev, 2023-03-24, about whether AI tutors in ChatGPT plugins will revolutionize education

milemi, 2023-03-25, about whether GPT-4 beating humans at coding competition tasks means it can really program

yamrzou, 2023-03-26, about an essay arguing that ChatGPT really understands, despite skeptics

Alex3917, 2023-03-26, about whether AI-generated marketing campaigns and emails are impressive or just generic

topper-123, 2023-03-26, about a GPT-4 summary of Charlie Munger's views on moats and whether its quotes are real

OutbackWizard, 2023-03-26, about a GPT-4 reverse Turing test and how well ChatGPT mimics humans

civilized, 2023-03-26, about whether ChatGPT understands math, prompted by the essay "And yet It Understands"

snickerbockers, 2023-03-26, about whether ChatGPT should hand off tasks like arithmetic to specialized tools

rglover, 2023-03-26, about whether LLMs are just word predictors or fundamentally different from humans

osigurdson, 2023-03-27, about whether tech layoffs are driven by hopes that AI will replace employees

smt88, 2023-03-27, about whether LLMs like ChatGPT can build non-trivial things, after a Wozniak AI-anxiety interview

lisasays, 2023-03-28, about whether LLM translations are good enough to use without human correction

visarga, 2023-03-28, about whether LLMs that can write and train new LLMs are effectively self-replicating

quickthrower2, 2023-03-29, about whether learning a programming language still makes sense given AI

spaceman_2020, 2023-03-30, about whether LLMs will soon autonomously design, code and distribute whole apps

alienicecream, 2023-03-31, about whether language models trained to mimic text can achieve insight or general AI

deafpolygon, 2023-04-06, about whether AI will replace software engineers, given the job is mostly communication

cornholio, 2023-04-11, about AI agents simulating human behavior and telling simulacra from sentient AI

NickC25, 2023-04-12, about an AI voice-cloning kidnapping scam and whether we need ways to verify real humans

The_Colonel, 2023-04-14, about whether GPT can catch its own logical fallacies, given Brandolini's law

RandomLensman, 2023-04-20, about whether ChatGPT can reason and apply knowledge, or is nowhere near AGI

crakenzak, 2023-04-24, about whether LLMs learn new facts at inference time or lose them beyond the context window

celdon25, 2023-04-26, about whether AI is fundamentally defective until it can say where it learned things

celdon25, 2023-04-26, about whether LLMs can explain their reasoning and cite sources the way an AGI should

srslack, 2023-05-01, about whether GPT-4 running under AutoGPT could take control, after Hinton's danger warning

PaulDavisThe1st, 2023-05-01, about whether LLMs really reason when they solve math word problems

nradov, 2023-05-01, about whether Microsoft's "Sparks of AGI" paper shows GPT-4 nearing general intelligence

ux-app, 2023-05-02, about the commenter's bar for a tech singularity: an AI that cleans a toilet

danShumway, 2023-05-02, about whether prompt injection matters compared to aligning far more powerful future AI

ckemere, 2023-05-03, about DeepMind's CEO saying some form of AGI is possible within a few years

canadiantim, 2023-05-06, about an AI tool that makes realistic architectural renders from AutoCAD and Blender

phillipseamore, 2023-05-11, about a live AI-based global weather forecast shown on Show HN

kmac_, 2023-05-13, about whether large language models show emergent intelligence or just statistics

agentultra, 2023-05-16, about whether next-token prediction gives LLMs real mathematical reasoning

simonh, 2023-05-16, about whether the field now requires an embodied Turing test for AI

alienicecream, 2023-05-26, about whether GPT-4 really counts or just outputs plausible-sounding answers

LightBug1, 2023-05-28, about whether the commenter would let a car drive itself if it drove better than them

endisneigh, 2023-06-06, about tips for coding with ChatGPT, and whether writing code is the real bottleneck

zirgs, 2023-06-11, about whether generative AI makes learning programming or music pointless

XorNot, 2023-06-14, about whether training AI on filtered synthetic data could produce better models

tarmol, 2023-06-15, about a veteran engineer's fear that GPT-4 threatens their job security

timacles, 2023-06-15, about whether AI will replace programming languages and write all code within 20 years

8chanAnon, 2023-06-17, about fatigue with AI chat products and their tendency to hallucinate

tremon, 2023-06-22, about testable hypotheses for how LLMs differ from human minds

abhaynayar, 2023-06-30, about whether worries about hyperpersuasive AI are justified

Tainnor, 2023-07-03, about whether GPT-4 writing working scripts means programmers won't write code in 10 years

Slartie, 2023-07-04, about whether GPT-4 writing scripts means people will stop writing code within ten years

sillysaurusx, 2023-07-07, about Mechanical Turk workers using AI to automate their tasks

aurareturn, 2023-07-10, about how well Copilot works in daily coding and how much context it uses

bumby, 2023-07-11, about whether LLMs' weaker performance on counterfactual tasks like base-8 arithmetic shows a lack of reasoning

PrimeMcFly, 2023-07-14, about whether AI-generated entertainment will shrink the movie and TV industries

villgax, 2023-07-19, about an Nvidia driver release and what would truly test LLM coding

thesuperbigfrog, 2023-07-20, about Google testing an AI tool called Genesis that writes news articles

Shawnj2, 2023-07-22, about whether Hollywood writers striking over AI are right to fear replacement

13years, 2023-07-27, about whether banning ChatGPT in schools is like banning calculators in the 80s

hardware2win, 2023-08-11, about an essay arguing AI will not undergo a sudden hard takeoff

skepticATX, 2023-08-11, about how we could tell that artificial general intelligence has been reached

wizofaus, 2023-08-11, about what would count as AGI, such as learning to drive like a human

worrycue, 2023-08-25, about whether LLMs reason or only memorise, given that they hallucinate

Art9681, 2023-08-26, about whether cheap open coding models like CodeLlama threaten software engineering jobs

saurik, 2023-08-27, about whether critics of AI are moving the goalposts on what counts as intelligence

kortilla, 2023-08-29, about whether ChatGPT Enterprise output can be trusted given hallucinations

crazygringo, 2023-09-03, about why AI can't replace actors, as SAG-AFTRA's strike expands to video games

mlhpdx, 2023-09-11, about whether Project Gutenberg's AI-narrated audiobooks are high quality

nradov, 2023-09-18, about whether AI can replace human fighter pilots, after a pilot ejected from an F-35

modeless, 2023-09-19, about Bard's new ability to search a user's Gmail and Google Drive

ryandrake, 2023-09-20, about AI generating new episodes and albums of discontinued shows and bands

kevinventullo, 2023-09-21, about whether algorithmic discoveries in mathematics count without proofs or understanding

WendyTheWillow, 2023-09-25, about whether ChatGPT's new voice and image features threaten software developers' jobs

xbpx, 2023-09-29, about a guide to web scraping with ChatGPT

karmakaze, 2023-10-03, about Bing Chat being tricked into solving a CAPTCHA with a dead-grandma locket story

joegibbs, 2023-10-05, about whether coding will still exist as a profession in five years

Sakos, 2023-10-16, about GPT-4V reading an image whose "attack while its tail is up" advice is wrong

kridsdale1, 2023-10-18, about whether AI can animate human movement without falling into the uncanny valley

__loam, 2023-10-28, about whether LLMs trained on books could replace reading them, for copyright purposes

nradov, 2023-11-01, about whether AI could replace airline pilots, especially in emergencies

dom96, 2023-11-05, about ChatGPT defeating email address obfuscation on websites

SkyMarshal, 2023-11-09, about an AI tool that summarizes arXiv papers and whether abstracts already do that

Barrin92, 2023-11-14, about whether ChatGPT counts as a general system in a paper on levels of AGI

thaanpaa, 2023-11-14, about whether AI coding tools mean the end of programming as a craft

dartos, 2023-11-14, about an LLM-based agent that performs hundreds of tasks in Minecraft

nradov, 2023-11-17, about whether LLMs could maintain legacy software such as the Pentagon's accounting systems

morsecodist, 2023-11-18, about whether GPT-4 is already AGI with human-level reasoning

ric2b, 2023-11-18, about whether GPT-4 is already AGI rather than a parroting machine

scarmig, 2023-11-18, about whether an AI must prove something like the Riemann Hypothesis to show understanding

civilized, 2023-11-18, about what separates humans from LLMs in Hinton's "glorified autocomplete" debate

Eggpants, 2023-11-18, about Geoff Hinton's view that LLMs are more than glorified autocomplete

ajuc, 2023-11-21, about whether AI imitating an artist's style is comparable to human artists studying it

hn_throwaway_99, 2023-11-23, about GPT-4 Turbo's better multi-digit arithmetic and whether LLMs truly reason about math

lamontcg, 2023-11-23, about reports that OpenAI's Q* model solved grade-school math problems

gryfft, 2023-11-23, about whether OpenAI's economic-value definition of AGI is a fair measure of intelligence

wbhart, 2023-11-23, about whether LLMs plus backtracking search could prove mathematical theorems, after OpenAI's rumored breakthrough

pegasus, 2023-11-23, about whether generative AI can produce consistent movie-quality video of 3D worlds

trashtester, 2023-11-24, about whether the first self-improving AI would gain an unbeatable lead

anonylizard, 2023-11-27, about a $10M prize for an AI winning gold at the math olympiad

still_grokking, 2023-11-29, about AI dev tools at Google that reportedly generate more code than humans

still_grokking, 2023-11-29, about whether AI tools can safely delete and simplify code in large codebases

pixl97, 2023-11-30, about whether seemingly small problems like self-driving require general intelligence

KPGv2, 2023-12-02, about whether a real-time "Babel Fish" translator is feasible, following Meta's Seamless speech models

Eridrus, 2023-12-04, about whether GPT-4 is useful for writing AWS IAM policies, under an IAM tool launch

nradov, 2023-12-05, about whether LLMs will take over most of the mental work done by doctors

dogmatism, 2023-12-05, about whether LLMs will take over most of the mental work doctors do

modeless, 2023-12-05, about whether LLMs can help with diagnosis when patient data is scattered across offices

galaxyLogic, 2023-12-07, about Gemini Pro failing a puzzle about how many sisters a girl has

Jensson, 2023-12-10, about which tasks LLMs still fail more often than average humans do

rmbyrro, 2023-12-12, about whether AI will make the supply of software effectively infinite

cornel_io, 2023-12-13, about whether AI could start a self-improving chain reaction, like nuclear fission

tim333, 2023-12-13, about Tesla's history of full self-driving promises and whether only edge cases remain

zer0c00ler, 2023-12-14, about Andrej Karpathy's view of LLM hallucinations and how far to trust LLM output

samirillian, 2023-12-14, about an essay by an AI skeptic admitting they were wrong

eastof, 2023-12-14, about whether AI meets the definition of AGI as doing anything a human can do

taylodl, 2023-12-18, about whether AI is on a trajectory to surpass human intelligence

Buttons840, 2023-12-24, about whether the Turing test turned out to be a bad metric for AI

mattdesl, 2023-12-28, about the NYT lawsuit showing GPT can repeat copyrighted articles word for word

brucethemoose2, 2024-01-02, about how generative AI threatens the income of furry artists

pjmlp, 2024-01-02, about whether LLMs had made programmers obsolete by 2024, as some feared

int_19h, 2024-01-05, about whether open models rival GPT-3.5, after an amateur's 7B model topped leaderboards

golol, 2024-01-08, about what definition of AGI computers haven't already met

CuriouslyC, 2024-01-20, about AI researchers expecting AI-generated fakes to become undetectable

londons_explore, 2024-02-08, about whether AI can spot mistakes in its own training data

choilive, 2024-02-15, about Gemini 1.5's 10M-token context window and whether huge contexts are useful

ThePhysicist, 2024-02-19, about how people with mainly digital creative skills should plan for AI replacing them

xnorswap, 2024-02-19, about why asking LLMs to generate prime numbers is a useful test

chx, 2024-02-19, about whether LLMs reason or just do statistical optimization, given their gullibility

gfodor, 2024-02-21, about whether AI models like Sora training on art should be treated like human learners

wakawaka28, 2024-02-26, about whether AI could cheaply rewrite legacy software in memory-safe languages

dmitrygr, 2024-02-26, about what would count as a meaningful step towards AGI

kjreact, 2024-02-27, about whether anyone has built a truly fully autonomous car

onlyrealcuzzo, 2024-02-28, about what AI must do before most human work is unnecessary

magic_hamster, 2024-03-02, about whether AI-generated games will make learning game programming pointless

magic_hamster, 2024-03-02, about whether AI will generate complete movies and games within two years after Sora

bufferoverflow, 2024-03-05, about whether LLMs should hand arithmetic to calculators or get smart enough to do it

gtirloni, 2024-03-05, about whether generative AI lets developers skip learning the skills they are tested for

kromem, 2024-03-12, about how close AGI is, and how its definition changes the answer

IncreasePosts, 2024-03-12, about whether AI getting top scores on AP tests makes it AGI

HarHarVeryFunny, 2024-03-12, about Devin, the AI software engineer, and whether AI can replace developers

HarHarVeryFunny, 2024-03-13, about whether AI can do a developer's whole job, not just coding

huytersd, 2024-03-13, about whether the Figure 01 robot demo is real-time LLM control or predefined actions

Retr0id, 2024-03-16, about whether language models like ChatGPT can understand or generate images and ASCII art

seccode, 2024-03-17, about having AI chatbots ask clarifying questions instead of just answering

bsder, 2024-03-18, about whether AI-generated code is maintainable enough for real software development

pants2, 2024-03-20, about whether any AI can pass as human on a live video call

dingnuts, 2024-03-21, about whether Vernor Vinge's prediction of superhuman intelligence within thirty years has come true

YetAnotherNick, 2024-03-24, about AI assistants on phones acting as agents across apps

Jensson, 2024-03-25, about whether code execution tools count as an LLM's "pen and paper" for reasoning

dingnuts, 2024-03-28, about ChatGPT making up code for a small, obscure programming language

anon373839, 2024-04-04, about whether fine-tuned AI models will threaten lawyers' work

reacharavindh, 2024-04-07, about a local LLM app for asking questions about uploaded documents

breadwinner, 2024-04-08, about Musk predicting AI will be smarter than the smartest human by next year

anonylizard, 2024-04-13, about whether Gemini 1.5 reading Harry Potter proves anything, since the books are in training data

thefourthchime, 2024-04-14, about comparing GPT-4, Claude and Gemini on the same prompts

aleph_minus_one, 2024-04-18, about a NYT claim that today's AI systems pass the Turing test easily

nicklecompte, 2024-04-18, about a New York Times claim that today's AI passes the Turing test with flying colors

devilbunny, 2024-04-20, about whether AI could replace doctors who rely on intuition, such as anesthesiologists

fragmede, 2024-04-21, about whether robots will be able to repair each other

Vegenoid, 2024-05-13, about whether an AI that fools most people in conversation counts as human-level AGI

JimmyRuska, 2024-05-15, about how 1980s logic-based AI like Prolog compares with today's LLMs

borgdefense, 2024-05-15, about whether needle-in-haystack tests measure real understanding of long documents

belter, 2024-05-15, about whether AI art is creative, given that humans create without huge datasets

autoexec, 2024-05-17, about Toon3D and whether AI could let fans rewatch cartoons from new perspectives

Sharlin, 2024-05-18, about whether LLMs trained on books are like compressed copies of copyrighted works

codexon, 2024-05-18, about whether LLMs that score well on reasoning benchmarks actually understand logic

lukan, 2024-05-24, about whether an LLM answering specific questions about a subject shows understanding

l33tman, 2024-05-25, about whether ChatGPT's hallucinations are acceptable when it is close enough to correct

WalterBright, 2024-06-09, about a claim that LLMs will never predict chaotic systems

PaulHoule, 2024-06-10, about a summary of the "Situational Awareness" essay predicting rapid AI progress

ben_w, 2024-06-10, about whether there is evidence of human-level AI coming soon, amid ChatGPT-4o hype

jerf, 2024-06-12, about whether GPT is useful for high-level math, given Terence Tao's experiments

proc0, 2024-06-13, about how to describe AI progress clearly without the vague term AGI

Affric, 2024-06-14, about whether Aschenbrenner's prediction of AGI within a few years is naive

therobots927, 2024-06-14, about whether AI will upset most of the economy within 20 years

lassoiat, 2024-06-18, about GPT-4o reaching 50% on ARC-AGI and whether LLMs are approaching AGI

kaba0, 2024-06-20, about whether children are smarter than the AI meant to protect them from online grooming

poincaredisk, 2024-06-21, about how hard writing a working exploit for an XNU kernel buffer overflow really is

deeth_starr_v, 2024-06-21, about whether AI is already replacing junior developers, in the context of an AI bubble

sph, 2024-06-21, about whether professionals who refuse AI tools will fall behind

kiernanmcgowan, 2024-06-21, about whether ChatGPT fails to understand causation

xorcist, 2024-06-26, about whether LLMs can analyze government data by writing programs to process it

tim333, 2024-07-06, about whether AI could become superintelligent and make new scientific discoveries

soloist11, 2024-07-06, about whether AI can formalize mathematics like algebraic topology in Lean

xracy, 2024-07-07, about whether LLMs that fail simple questions can help with real science

Pinkthinker, 2024-07-12, about OpenAI's five-level roadmap from current AI to AGI

rolph, 2024-07-14, about companies bolting AI onto products where a basic algorithm would do

abernard1, 2024-07-20, about whether LLMs encode conceptual models in their intermediate-layer parameters like brains do

disgruntledphd2, 2024-07-21, about using LLMs to convert R codebases that rely on statistical library functions

ccvannorman, 2024-07-21, about litmus tests for AI coding, like writing game graphics shaders

skyde, 2024-07-24, about whether LLMs should do arithmetic themselves or use Python tools

VincentEvans, 2024-07-24, about whether tokenization excuses LLMs failing to count letters in "strawberry"

imtringued, 2024-08-01, about whether user feedback could give OpenAI an experience flywheel to improve its models

alkonaut, 2024-08-03, about whether DARPA's TRACTOR can automatically translate C programs into Rust

gosub100, 2024-08-03, about a project to auto-translate C to Rust, and whether AI could write device drivers

reify, 2024-08-05, about an AI mental health companion app with real-time voice chat

hugs, 2024-08-08, about whether multimodal LLMs make vision-first UI automation practical

Guvante, 2024-08-10, about whether ChatGPT's confident, unsourced answers make AI a systemic risk