Sanadkii 2017, dowladda Shiinaha ayaa shaacisay qorshe ballaaran oo ay ku dooneyso inay hoggaanka u qabato horumarinta Teknolojiyada Caqliga Macmalka (AI) ilaa 2030. Qorshahaas wuxuu ujeeddadiisu ahayd in Shiinaha gaaro guulo waaweyn oo teknoolojiyada AI ah sannadka 2020. Si kastaba ha ahaatee, dhamaadkii 2022, shirkadda Mareykanka ee OpenAI ayaa soo saartay ChatGPT taasoo caalamka ka yaabisay islamarkaana Shiinaha ku qabsatay inuu laba jibaaro dedaaladiisa AI.
Muddadaas, shirkadaha waaweyn ee teknoolojiyada Shiinaha ayaa la daalaa-dhacayay cunaqabateyn ay dowladdoodu ku soo rogtay oo socotay 18 bilood, taas oo qiimo dhan $1 tirilyan ku dhintay qaybta teknoolojiyada dalka. Waxay ku qaadatay in ka badan sannad in dhowr chatbot oo Shiinees ah ay helaan ogolaansho rasmi ah si loo sii daayo. Waxaa sidoo kale jiray walaac laga muujinayay in hab-dhaqanka xayiraadda macluumaadka ee Shiinaha uu caqabad ku noqon karo qorshayaashooda horumarinta AI.
Dhanka kale, Madaxweynihu hore ee Mareykanka Joe Biden ayaa bishii Oktoobar 2022 soo rogay xeerar xakameynaya dhoofinta chips-ka casriga ah ee loogu talagalay barnaamijyada waaweyn ee AI. Tani waxay caqabad weyn ku noqotay rajada Beijing ee hoggaanka AI-ga sannadka 2030.
Eric Schmidt, oo ahaa maamulihii guud ee Google intii u dhexeysay 2001 ilaa 2011, ayaa wareysi uu siiyey Bloomberg ku sheegay in Maraykanku “aad uga horreeyo Shiinaha” dhinaca CaqligaMacmalka(AI).
Sababaha Shiinaha Ugu danbeeyo Tiknoolajiyadda AI
Schmidt ayaa soo bandhigay afar sababood oo keentay in Shiinuhu aan horumar weyn ka sameyn karinAI.
1. In shiinaha uusan heysan chip-yada muhiimka ah .
2. Xog yari dhinaca tababaridda qaabka luuqadaha(LLM) weyn.
3. Hoos u dhac weyn oo ku yimid maalgashiga shisheeye
4. Shiinaha oo Diiradda saaray shirkadaha abuuraya barnaamijyada faa’iido-doon ah, halkii ay ka horumarin lahaayeen aaladaha AI-ga ah.
Balse Eric Schmidt, oo markii hore aaminsanaa in Maraykanku leeyahay hoggaanka AI sanado badaana ka horreeya Shiinaha ayaa badalay fikirkiisi isagoo bishii Nofeembar 2024 lix bilood kadib wareysigis hore kadib sheegay in horumarka Shiinaha uu la yaab ku noqday.
shiinahana yahay wadan AIgiisu xowli ku socdo.
Mareykanku wuxuu ku tiirsan yahay shirkadaha Nvidia iyo AMD si loo soo saaro chips-ka casriga ah ee loo adeegsado AI. Si kastaba ha ahaatee, Shiinaha ayaa la sheegay inuu keydiyay chips-kii ugu dambeeyay kahor inta aan cunaqabataynta lagu soo rogin. Wararka kale waxay tilmaamayaan in qaar ka mid ah chips-ka la mamnuucay ay si sharci-darro ah ku galeen gudaha Shiinaha.
Inkasta oo xayiraadahaas ay sii adkeeyeen sannadihii la soo dhaafay, haddana Shiinaha ayaa ku guuleystay inuu horumar ka sameeyo moodallada AI isaga oo isticmaalaya chips-ka awoodooda hooseysa. Shirkadda Tencent ayaa soo saartay moodallo heer sare ah oo ku shaqeeya chips-ka aan mamnuuca ahayn.
Shirkadda DeepSeek ayaa ku qarashgareysay 6 milyan oo doolar si ay u tababarto moodalkeeda cusub, halka Meta ay ku bixisay lacag dhan 60 milyan oo doolar. ChatGPT-na waxaa ka baxday 100 milyan oo doolar. Qiimaha yar ee shirkadan ku soo saartay horumarka AI ayaa ka yaabisay khubarada AIga sida ay shirkadaan u gaartay horumar weyn iyadoo isticmaalaysa maalgelin kooban.
DeepSeek waxaa aasaasay Liang Wenfeng bishii Luulyo 2023. Liang oo hore uga soo baxay Jaamacadda Zhejiang ayaa sidoo kale aasaasay shirkad maalgashi oo la yiraahdo High-Flyer sanadkii 2015. Liang wuxuu leeyahay riyo ah in uu dhiso AI guud (AGI) oo ka sarreeya ama la siman aadanaha marka la eego hawlaha badan.
Shirkadda DeepSeek waxay soo saartay model cusub oo caqli-garashadiisa la barbar dhigi karo ChatGPT o1. Modelkan oo la yiraahdo DeepSeek R1 ayaa noqday mid xiiso badan oo si weyn looga hadal hayo Caalamka.
DeepSeek oo ah shirkad Shiinees ah oo ku taalla Hangzhou, ayaa sheegtay in R1 uu la siman yahay ama xitaa ka wanaagsan yahay ChatGPT o1 marka la eego dhowr waxyaabood oo muhiim ah, iyadoo ku shaqeynaysa kharash aad uga hooseeya marka loo eego shirkadaha AI ee Mareykanka.
DeepSeek waxay guusheeda gaartay iyadoo ay Maraykanka aad u adkeeyee.cunaqabateynta ee ku saabsan chip-yada casriga ah ee loo dhoofiyo Shiinaha. Halkii ay cunaqabateyntu hoos u dhigi lahayd awoodda AI ee Shiinaha, waxay keentay in ay soo.baxaan shirkado AI sida DeepSeek ayna horumar sameeyaan iyagoo diiradda saaraya wax ku oolnimo, wadaagida ilaha, iyo iskaashi.
Modelka R1 ayaa loo sameeyay qaab tababar cusub oo hoos u dhigaya culeyska GPU-yada ay isticmaalayaan, kuwaas oo xawaarahooda la xadiday si ay uga u qabtaan howlo la mid ah chip-yada ugu sarreeya ee Nvidia.
Waxqabadka R1
Modelkan ayaa si gaar ah loogu amaanay awooddiisa xallinta hawlaha adag, gaar ahaan kuwa xisaabta iyo barnaamijyada. Wuxuu adeegsadaa habka loo yaqaan “chain of thought” oo u oggolaanaya inuu tallaabo-tallaabo uga shaqeeyo su’aalaha adag.
Dimitris Papailiopoulos oo ka tirsan Microsoft AI Frontiers ayaa sheegay in waxa ugu yaabka badan ee uu arkay ay tahay fudaydka farsamada ee R1. DeepSeek ayaa sidoo kale sii deysay lix nooc oo yaryar oo ka mid ah R1 kuwaas oo si toos ah looga shaqeyn karo kombiyuutarrada laptop-ka.
Shirkadaha Shiinaha ayaa sidoo kale ku dhiirranaya mabda’a furfurnaanta (open-source). Alibaba Cloud ayaa sii daysay in ka badan 100 model oo cusub oo furan oo taageera 29 luqadood. Sida laga soo xigtay warbixin ay soo saartay Akadeemiyada Teknolojiyadda Isgaarsiinta Shiinaha, 36% model-ka caalamiga ah ee AI waxay ka yimaadeen Shiinaha, taas oo ka dhigaysa inay kaalinta labaad ka galaan adduunka.
DeepSeek-R1 waxa uu ka jaban yahay ChatGPT R1, iyadoo ChatGPT uu ku kacayo $15 iyo $60 halkii milyan ee tokens-ka gelinta iyo soo-saarka, halka DeepSeek-R1 uu qiimo ahaan u dhigmo $0.5 iyo $2.19. Tani waxay sababtay in DeepSeek-R1 uu halmar gilgilay saamiga suuqyada AI ee Mareykanka iyo Reer galbeedku hoosna u dhigay dhaqaalaha.
DeepSeek-R1 waxa uu qalqal geliyay shirkadaha waaweyn ee Mareykanka, waxaana halista kale ee weyn ee ay dareemayaan reer galbeedku ay tahay in DeepSeek uu yahay mid furan (open source), taas oo micnaheedu yahay in dhammaan dadka daneynaya ay wax ku dari karaan isla markaana ay heli karaan macluumaadka AI-ga. Sidoo kale, dadka maalgashadayaasha ayaa arki kara fursado cusub, taas oo daaha ka qaadaysa sirta iyo monolopoliga uu Mareykanka AI ku heysto.
Horumarka DeepSeek ayaa soo jiitay taageero weyn oo ka timid bulshada caalamiga ah ee cilmi-baarista AI, iyadoo shirkaddu si xor ah u wadaagtay hal-abuurnimadeeda si ay u soo jiidato isticmaaleyaal iyo tabarucayaal cusub. Xeeldheerayaasha ayaa rumeysan in guusha shirkaddan ay shakki gelin karto istiraatiijiyadda Mareykanka ee xayiraadaha tiknoolajiyadda AI, maadaama DeepSeek ay muujisay in guul weyn lagu gaari karo iyada oo aan la isticmaalin chip-yada casriga ah ee reer Galbeedka.
Haddaba, DeepSeek ayaa noqotay tusaale cad oo muujinaya sida hal-abuurnimada iyo hufnaanta tiknoolajiyadeed ay ka sareyn karto xayiraadaha caalamiga ah, taas oo Shiinaha ka caawisay inuu ku sii xoojiyo booskiisa warshadaha AI ee caalamka.
W/Q Claahi Daahir Duqow
