OpenAIandAltmanareon a questtodevelopAGIorartificialgeneralintelligence. Thecompanybehindthepopularchatbot, ChatGPT, OpenAIwasfoundedin 2015 by a numberofresearchers, academicsandentrepreneurs, includingSamAltman, GregBrockmanandElonMusk.
Theiroriginalmissionstatementwassomethingliketoadvanceartificialintelligencein a waythatbenefitsallofhumanity, unconstrainedby a needforfinancialreturn.
It's likeyouhaving a studentthatisputinto a classroomandthenyougiveit a bunchofinformationandthenyouhavetotryandanswer a bunchofquestions. Andbaseduponwhetheritanswersthosequestionscorrectlyornot, yougiveitpositivefeedbackornegativefeedback.
Soessentially, whatitboilsdowntoisit's a wayforpeopleorsystemstofurtheriterateon a modeltodirectittowards a certainsetofbehaviors. In 2016, OpenAIalsoreleasedUniverse, a tooltotrainintelligentagentsonwebsitesandgamingplatforms.
ChatGPTis a chatbotpoweredbyartificialintelligence, whichbasicallymeansitingests a tonofhumanlanguagefromalloverthe
ChatGPTは人工知能を搭載したチャットボットです。
Internet, maybeReddit, Googlesearchresults, maybetravelblogs, whateveritwastrainedon, itwilluseallofthattocomeupwith a responsetoyouthatsoundslike a way a humanwouldansweryourquestion. Here's BrianChesky, a closefriendofAltmanandCEOof
Airbnb, recountingthemomentChatGPTlaunchedduring a recentinterview. Itwas a phenomenonunlikeanythingwe'd seenprobablysincethelaunchoftheiPhone. I havenorecollectionofanything.
Chatbotshavebeenaroundfor a while, but I thinkwhatsetChatGPTapartherewasthefactthatitdebutedat a timewhenpublicinterestandcuriosityaroundAIwasatanall-timehigh.
Andthentheybuild a construction, a modelofhowthatallrelatestoeachother. Thereasonthatwecallthemlargelanguagemodelsisbecausethere's somuchdataintherethatitisjustordersandordersandordersofmagnitudebiggerthanthelanguagemodelsthatwe'vehadinthepast.
InJuly, OpenAIannouncedthatitwastesting a newsearchenginetoincorporateintoChatGPTthatwouldgiveusersfastandtimelyanswerswithclearandrelevantsources.
7月、OpenAIはChatGPTに組み込む新しい検索エンジンをテストしていると発表した。
Thetechnologyisseenas a directcompetitortoGoogleSearch.
この技術はグーグル検索の直接のライバルと見られている。
Whatwe'veseenfromOpenAIoverthepastcoupleofyearsisthattheydefinitelyseemtohaveaninterestinbeing a commercialcompanyandinbeing a companythatservesthelargestbusinessesintheworld, thatservesenterprisecompanies.
Thereare a lotofbiotechresearchapplications, drugdiscovery, certainapplicationswithinlegalservices, anythingthatrequiresingesting a bunchofdataordocumentsorinformationandthensummarizingit, creatingtasksbasedoffofit, creatingkeytakeaways.
Ifwelookatthesoftwaredevelopmentspace, I thinkthatthatistheareathathasbeeninfluencedprobablymostobviouslyatthispointintime. A recentreportfoundthatjobsinthebankingsectorwerethemostlikelytobeimpactedbyAI, followedbyjobsintheinsuranceandenergysectors. Asforwhatthatwillmeanforhumanworkersinthesesectors, Curransaysitmaystillbetooearlytotell.
Aswitheverytechnologicalinnovationanddevelopment, there's goingtobe a shiftinwherethejobsareandtherewillbesomejobloss.
あらゆる技術革新や開発と同様に、雇用の場が変わり、雇用が失われることもあるだろう。
We'regoingtoseesomejobtransformation, but I thinkthatthenetgrowthisgoingtobesignificantlygreaterthanthejobsthatarelostatthispoint. Allthispotentialhasgotinvestorsandtechgiantsexcited, manyofwhichareheavilyinvestinginAIstartupsaswellasbuildingouttheirowntechnology.
ButinearlyAugust, hefiled a similarlawsuitagainstAltman, BrockmanandOpenAI, allegingthathewasmanipulatedintoco-foundingOpenAIonthebasisthatitwouldbe a nonprofit.
Muskhas a competingAIstartupknownasXAIand a chatbotcalledGrok.
マスクはXAIというAIスタートアップとGrokというチャットボットを競合させている。
This, despiteMusksigning a letterlastyearalongsideover a thousandothertechleadersandresearcherscallingfor a moratoriumonAIdevelopmentbecauseofprofoundriskstosocietyandhumanity.
So I'veseenestimatesbetween a fewmillionandonebilliontotrainanAImodellikethisandtorunthemdaybydaycouldbehundredsofthousandsormore. There's a lotofconcentratedpowerrightnowbetweenstartupsandbigtechcompanies. AndwemayseethatshiftdependingontheresultsoftheFTCinquiryandotherregulatoryscrutinyhappeningrightnow.
Butthat's where I thinkwe, societyandgovernments, have a roletosay, youknow, we'llallowsomeusecasesoftechnologywe'renotcomfortablewith, butinsomeplaceswearegoingtodraw a line.
InJuly, OpenAIalsoreassigneditstopAIsafetyexecutiveto a rolefocusedonAIreasoning. However, OpenAIdoescontinuetousewhattheindustrycallsredteamers. Redteamerstest a company's modelsfromwithintoseeexactlyhowdisturbing, unsafe, negativeordangeroustheycanbeandreportthattothecompanysothathopefullytheycanputupmoreguardrailsandstopthatfromhappening. Sincethebeginning, copyrightissueshavealsoplaguedOpenAI.
Post, BusinessInsiderandTime, thesecompaniesaregettingpaidtosharetheircontentandarchiveofcontentwiththecompanytohelptrainitsmodelsandtoshowupwithinChat2BTasanswerstouserquestions.