Ngo-2014, i-Google yasungulwa I-I.I.E.I.E.I. isekelwe eLondon njenge-start-up yokufaka emakethe ye-Machine Learning, ebizwa nge-availability ephakeme ye-computing eyenziwe ngu-Moore's Law. Ngezinyanga elizayo, ngo-2015, i-partnership ephakeme phakathi kwezinye ama-investors eSilicon Valley, kuhlanganise u-Peter Thiel no-Elon Musk, yasungulwa i-OpenAI, i-AI laboratory eyenziwe ngempumelelo. Okokuqala, Amazon, ngemva kweminyaka eminingi yokuvimbela i-inthanethi ebizwa ngokuthi i-AI phakathi kwezinhlayiyana ze-tech powerhouses, waya umkhakha we-AI kanye ne-AI Ethics research lab Anthropic ngokusebenzisa i-$4 billion investment kanye ne-computing partnership. I-US $ 500 million i-akquisition ye-DeepMind I-partnership ye-Microsoft ngo-2019 Yenza i-Alliance ye-Strategic I-US $ 500 million i-akquisition ye-DeepMind I-partnership ye-Microsoft ngo-2019 Yenza i-Alliance ye-Strategic Phakathi neminyaka embalwa yokufika kwe-AI emakethe yobuchwepheshe, izixazululo zayo ekhompyuthazweni zithunyelwe ekuthunyelwe kwebhizinisi zezifundo kanye nesikhathi esincane yokuthengisa imikhiqizo ye-AI. Lezi zophando ezivamile, lapho amabhizinisi ze-AI baye zihlanganisa ukuthuthukiswa okwenziwe ngempumelelo ngokufaka nezifundo ze-universities kanye nokuthuthukiswa ngokushesha kwegama le-machine learning, zihlukile ngokushesha futhi ngokushesha ngenxa yokuthuthukiswa okusheshayo kwe-OpenAI ngenxa yokuthuthukiswa kwe-ChatGPT nge-internet ekupheleni kwe-2019. Yini elandelayo kuyinto isikhathi esizayo yokudluliselwa kwamandla, ukuguqulwa kwezinhlelo, Ukubonisa isigaba Ukuqala ukuhlaziywa kwami, ngithanda ukuthi kunezinto ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi, noma ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi ezingu-inthanethi-inthanethi, kukhona inani elingu-inthanethi elingu-inthanethi-inthanethi-inthanethi ezingu-inthanethi-inthanethi-inthanethi-inthanethi ezingu-inthanethi-inthanethi-inthanethi-inthanethi-inthanethi-inthanethi Ngokuvamile, eminyakeni noma amabili, i-OpenAI iholela ibhizinisi le-AI, ukuguqulwa kwama-iterations ezintsha ze-ChatGPT ezivela-ke-back-to-back, zonke zithembisa izinzuzo ezingaphezu kwama-laicism ne-professional, ibonise i-OpenAI izinzuzo eningi njengokufaka isisekelo esikhulu se-consumer kanye nokuvimbela okuphakeme kwebhizinisi se-for-profit. Ngokuvamile, i-Google, inkampani ye-tech nge-cloud computing system eningi kakhulu kanye nama-massive amasethi ye-financial ne-human capital resources, akazange uvumela i-chatbot yokukhuthaza i-ChatGPT kuze ku-2023, cishe unyaka ngemuva kokuph Nokho, njengoba kulokhu, amazwe amaningi ze-AI zihlanganisa ucwaningo, futhi amabhizinisi amabhizinisi amabhizinisi we-Google (i-Google Brain ne-Google DeepMind) zihlanganisa ngokufanelekileyo emaphrojekthi eziningana. Ngaphandle kwe-ChatGPT, i-ChattGPT inesibopho esigcwele yokushintshwa kwama-models ezintsha, zokusebenza kakhulu kubasebenzisi ngexesha lokushicilela. Ngakho-ke, akugcwalisa ukuthi nangaphandle kwe-R&D kuya ku-commercialization (i-Google Brain wahlala ukuthuthukiswa kwe-Architecture ye-Transformer AI, isikhunta se-LLMs ezifana ne-ChatGPT; i-Google DeepMind ikakhulukazi wahlala i-Nobel Prize for its research in using Reinforcement Learning to develop protein folding predictions using the AlphaZero RL framework), i-Google's top brass eyenza izindlela eziningana nezinsizakalo zothumela i-Google (and the public, for that matter) ukuba Indlela Ngaba Ukungabonakaliswa kwe-OpenAI (ngokukwazi ukuthi i-Microsoft inesibopho le-AI emangalisayo), i-Google iye yenza izinguquko eziningana ukuhlinzeka izinhlelo zokusebenza zokukhiqiza. Ukusabela izindawo ezimbili ezidlulileyo, i-Google yasungulwa ngokuvamile ku-division eyodwa ngaphansi kwegama ye-DeepMind ngo-March 2023, okuholela Ukubuyekezwa ngokushesha kanye nokuhlanganiswa kwe-talent e-division ku-Bard team eyenziwe ngoku. Kubalulekile futhi ukwazi ukuthi ngezinyathelo ezaziwa kuphela ngenxa yokufinyelela kwe-Google ku-OpenAI emakethe ye-chatbot, kodwa futhi ngenxa yokukhula okuqukethwe kwe-idea ye-chatbot. Uma amamodeli ezidlulile (njenge-BERT noma i-Stabil Diffusion) zithunyelwe kakhulu (i-NLP ne-image generation ikakhulukazi), ukuthuthukiswa kwe-multimodality kanye nokuvumela i-MoE (I-Mixture of Experts) paradigm angakutholela i-Google ukunciphisa umqondo we-models angu-5 ezahlukile angakwazi ukwenza izinto angu-5 ngokuzimela umqondo we-model eyodwa enamandla angakwazi ukwenza izinto ezingu-5. merged the Google Brain and Google DeepMind teams Ukuqhathanisa Ukusebenzisa I-Google Brain kanye ne-Google DeepMind Ukuqhathanisa Ngokungafani ne-Google, i-OpenAI yakhelwe ngokuvamile ukubuyekeza ukubuyekeza imodeli yokuqala ye-AGI ehlabathini. Njengomzekelo, "ngokukwazi okuphumelela okuphumelela okuphumelela okuphumelela okuphumelela okuphumelela okuphumelela okuphumelela okuphumelela okuphumelela okuphumelela okuphumelela okuphumelela okuphumelela okuphumelela. I-Ideyali yaziwa elula: Njengoba abantu akuyona ngokuvamile akwazi ukuhlobisa ukusebenza kwe-labyrinthine ye-AI ye-supernatural, kuphela enye, engaphakeme kakhulu, i-AI angakwazi ukuhlola inkinobho yayo ephakeme kakhulu. I-Superalignment iqembu lwezinhlangano lwezinsizakalo yokuthuthukisa izinhlelo ezinikezele ukhuseleko kwe-AI, kodwa le nkqubo akuyona. Izinyanga ezidlulayo ziphumela ngezinga elikhulu sokuthintela ukuxhaswa phakathi kwe-OpenAI abacwaningi kanye ne-copper-top of the very much-for-profit organization, okuholela izinga ezihlukahlukene zebhizinisi ngaphakathi inkampani. Emva kokuphumula okuphumula kwe-OpenAI CEO Sam ngoMarch 2024 mayelana nokuhlukaniswa malunga ne-tempo lapho i-OpenAI ikhiqiza amamodeli. I-Superalignment Research Team Ukuhlobisa I-Superalignment Research Team Ukuhlobisa Nangona i-acme ye-disagreement eyenziwe ngezifundo malunga ne-profile mass resignation ye-Superalignment Team ngo-2024, amayuningi amayunithi eziyisisekelo mayelana nokuthuthukiswa okuphakeme nokuthuthukiswa kwe-AI amamodeli zangaphambili baye asebenzise inkampani. Ngo-2021, lapho i-OpenAI ngempumelelo ukwandisa ukuthuthukiswa kwe-ChatGPT-3 nge-series ye-rapid-releases, ezinye abacwaningi amakhulu (kuhlanganise uDaniel Amodei, owayemva kwe-VP ye-Research ku-OpenAI) baye baye baye baye zihlanganisa, Ngokuvamile, i-Anthropic, eyakhelwe njengelebhu le-research enikezela ukubaluleka kwezimo zokhuseleko ze-AI, ikakhulukazi ukucubungula, ku- state-of-the-art I-AI development, i-Anthropic ikakhulukazi ihlangene le ncebo, manje ngokuvamile ibhekwa njenge-champion eside ye-AI Ethics emzimbeni ye-AI yokushesha. Ukwakhiwa kwe-Rogue AI Lab Anthropic Ukwakhiwa kwe-Rogue AI Lab Anthropic Ukuqala Specialization Ukuqala ngo-May lokugqibela iye wathatha isinyathelo phakathi kwe-imeyili ye-Big Three (i-Anthropic, i-Google, ne-DeepMind) enikezela ekuphenduleni Ngo-climate ye-AI development. Njengoba kubhalwe ngaphambi, ingcindezi esivumelanayo eyakhelwe ukukhiqiza i-LLM ye-general-use eyenza izicelo ezininzi ngexesha. Njengoba izinzuzo ze-LLM zihlanganisa, ukucubungula kwangena, kuyoba ngokuvamile ukufinyelela i-AGI singularity ukuthi abantu abaninzi abalandeli ukufinyelela. I-counter-argument yami: isizwa ngokuthi i-singularity ngoba isizathu. Isiphumo efana nokwakhiwa kwe-Artificial General Intelligence kungenzeka ukuthi akuyona ngenxa ye-evolution of current Transformer-based architectures, okuyinto akudingeki ama-"building blocks" ye-intelligence yabantu (ngena: kinesthesis, ingcindezi ebonakalayo kanye ne- Ukuphakama kwe-Specialization Njengoba lo status quo eside ye-AI ukuthuthukiswa kwenziwa, kunjalo, kubonakala ukuthi kukhona umkhosi we-AI ephelele ukugcina indawo ye-# 1 ku-LLM izitolo ziye ziye ziye ziye ziye zithunyelwe kumadoda ngaphandle kwe-Google, umbhali we-Big Three nge-talent pool eningi, inethiwekhi ye-computing center eningi, futhi inani elikhulu le-data. Umkhosi we-tech, eyenziwe ngexesha elide ye-Google Bard, manje ibonise isakhiwo se-state-of-the-art ku-multimodal LLMs ngokuvamile ne-Gemini-2.5, isikhathi esizayo esizayo ku-2025 Google I/O kanye nesikhathi esizayo se-model, kuhlanganise i-SoTA entsha ye-veo3 Ngokuvamile, ngemuva kokuphuma kokuphuma kwe-User Testing ye-Gemini-2.5, i-OpenAI yasungulwa Ukuze ukuthenga i-AI hardware start-up io, isebenza ngama-designer e-Apple (ngokusho) ukuhlangabezana imishini yemvelo nge-creation kanye nokukhiqiza i-”AI Hardware” ezintsha. Kubalulekile ukuthi lokhu ukuxhaswa ukuthi i-OpenAI inikeza ukuxhaswa kwe-potencial market noma ukuxhaswa kwe-status ephelele kusuka ku-focus yayo ku-LLM ukuxhaswa ngokuvamile ngokuvumelana ne-Google's engaphambili emkhakheni. Ngitholile ukuthi kungcono ukuthi kuyinto emhlane, uma umfutho we-AI akuvumelekile ku-OpenAI ngempumelelo; kulesi mkhuba, angakwazi ukuxhaswa ne-AI inkampani ebonakalayo ngexesha ukuze zithembise ku-framework e Ukubuyekezwa I-$6.5 billion ibhizinisi Ukubuyekezwa I-$6.5 billion ibhizinisi I-Google inezinto zokuphendula imikhiqizo ye-side kanye nokugcina umphumela wayo ebuthakeme ku-AI race, futhi i-OpenAI yakamuva uqala indlela yokufinyelela ngokuvamile uma ufuna. Ngokungafani nabanye, i-Anthropic ibonise ngokuvamile ngokuvumelana ne-Claude, i-LLM yayo, yokukhipha, ngokuvumelana nokuvumela i-SoTA entsha nge-Claude Opus 4, nangokuthi umphumela wayo ku-benchmarking test kuyinto emangalisayo. Ngaphezu kwalokho, i-Anthropic ibonise ngokuvumelana ne-AI Ethics research organization eningi ehlabathini, ngokuvumelana nezinhlangano ezijoliswe ku-AI ukhuseleko kanye nezinhlangano zebhizinisi ze-int Ngokuyinhloko, lezi zihlanganisa izinzuzo zokusebenza kwe-AI zokusebenza kubalulekile ukuthi ekupheleni komhlaba jikelele, amabhizinisi amabhizinisi amabhizinisi angaphambili ziye zihlanganisa ukuthi zihlanganisa ngokushesha njengezimiphumela yemvelo ye-differentiation ye-product. Ungafakwa ukuthi uma i-Anthropic inikeza amamodeli ze-agent coding kanye ne-Google, ama-chatbots, ukufuna kwe-AGI ingatholakala ngokuvamile njengoba ukusetshenziswa kwamakhemikhali ebuthakathwa. Nokho, Ngitholile ukuthi ukucubungula ku-AI kunokwenzeka kakhulu ekugcineni, ngenxa yezizathu eziningana: 
 
 
 
 
 
 Competition breeds innovation. Even though additional resources are not being consumed by a single company in development, the lack of impetus to innovate would hinder advancements despite this fact. If Google already held an extremely stable monopoly, it would not have incentive to spend billions on additional R&D to upgrade a product already in high demand. 
 
 Companies do not share all discoveries. Assuming that all companies actually do invest into further development despite the emergence of specialization as mentioned above, AI companies have been historically reluctant to share the specifics of a model (even lobbying against proposed government legislation for 3rd party review of proprietary model information), which would practically negate any potential benefit from the “you do this, I do that” model of AI development. True open-source AI is still an ideal right now. 
 
 
 There is a heavy overlap across specialized models. For example, as video generation models get more and more convoluted, they require more complicated transformer models trained using extensive data, including the breakdown of textual prompts and inherent reasoning like that found in traditional LLMs (for example, a well-trained video-generator would inherently assume that a “wood board boat” would refer to a boat made out of wooden boards instead of a personified log boarding a ship). 
 
 This article is brought to you by Our AI, a student-founded and student-led AI Ethics organization seeking to diversify perspectives in AI beyond what is typically discussed in modern media. If you enjoyed this article, please check out our monthly publications and exclusive articles at  ! https://www.our-ai.org/ai-nexus/read Umhlahlandlela we-I.I. yethu, isakhiwo se-AI Ethics eyenziwe ngama-student eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi eyenziwe ngama-inthanethi ! https://www.our-ai.org/ai-nexus/read ... Imiphumela Ukuguqulwa kwebhulogi: Ukuphendula, ngokuvamile, ngenxa yokuba le mkhuba inesibopho ukuguqulwa kwendlu lokuthuthukiswa kwe-AI ngokuvumelana nathi. Kubandakanya izinhlamvu ukuthi angakwazi kuphela ukucacisa, ngenxa yokungabikho kwama-Connections e-Silicon Valley, kodwa kubandakanya konke nge-standard game theory. Thina ucabanga, ngokuvamile, ukuthi i-Google ne-Anthropic zombini zithunyelwe umfutho we-AI ngokuvumelana nokuthuthukiswa kwama-AI amamodeli zabo, i-Gemini ne-Anthropic. Nangona abacwaningi Uma akufanele ukuthatha indawo yayo, akufanele ukuba akwazi ukuchitha i-OpenAI, okuyinto ngokuvamile akuyona i-monopoly eningi ngeveki, okushisa izinzuzo zamabhizinisi amabili amabili ngokushesha. Yenza okufanayo kumabhizinisi amabili, futhi isampula esuka: Nangona kunjalo, uma usebenzisa le nkqubo kwinkampani ye-AI emangalisayo (ngokuthi, i-Mistral noma i-Meta), ungathola ukuthi kungcono kakhulu ukuthi le nkampani iya kuba i-AI superpower, njengoba kungcono kakhulu ukuchithwa kwimarike. Yintoni ukuguqulwa kwe-specialization ku-AI imikhiqizo i-pivotal moment ye-AI development? May Uma amabili abalandeli, i-company eyahlukile uya ku-dominant ngokuvamile. Njengoba kubangelwa ngaphambi, ukufuna kwe-AGI iyahlukile kwezinye izivivinyo ezivela ekuthuthukiseni imikhiqizo njengoba lokhu isibonelo yokuqala (eyazi) lapho imikhiqizo ye-prototype ingasiza ngokuqondile ekuthuthukiseni inguqulo esilandelayo. Lezi zimo ziya kubhalwe ngokuvumelana nokukhula kwe-Agentic AI, kodwa singatholakala isibonelo esiyingqayizivele kusuka ku-Google's AlphaEvolve model, isibonelo sokucubungula-and-coding eyenza ukwandisa umthamo we-DeepMind yokukhiqiza kanye nokwakhiwa kwe-AI nge-Google. I-framework ye-load distribution ye-computing centers ye-Google, okwandisa ukusebenza jikelele nge-1% okungenani (okungabikho kakhulu, kodwa kuncike uma usihlanganisa ukuthi zonke ama-computing centers e-Google zihlanganisa idatha nge-trillions ye-byte). Ngokusho ukuthi le nempendulo ivele (ie, amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizinisi amabhizin Ukuhlobisa Ukuhlobisa Ngokusho zonke lezi zincwadi, kungabikho ukuthi omunye ye-Triumvirate yokuthuthukiswa kwe-AI akugqoka emzimbeni yayo ngokushesha - zonke lezi zincwadi zithunywe emaphandleni kanye nezincwadi ze-venture capital, ukudluliselwa kwama-brand, futhi, okungenani kakhulu, izivakashi zomphakathi. Nge izigidi ze-financing zithunyelwe kumazwe ezingu-massive, sincoma ukuthi yini elungileyo lithunyelwe.

The is an opinion piece based on the author’s POV and does not necessarily reflect the views of HackerNoon.

Read My Stories

Lo msindo ukhiqizwa ngolimi lokuqala lwendaba!

We're at a fork in the way for AI Development

About Author

IMIBONO

HANG TAGS

LESI SIHLOKO SETHULWE NGAPHAKATHI

Related Stories

The Billionaires F*cked Up

Zulu Network: Moving The Bitcoin Economy Forward With a Two-Tiered Bitcoin Layer 2 Architecture

The Billionaires F*cked Up

Zulu Network: Moving The Bitcoin Economy Forward With a Two-Tiered Bitcoin Layer 2 Architecture

Light-Mode

Classic

Newspaper

Dark-Mode

Neon Noir

Minty

HN StartUps