Ishadi likaGartner lifana nombukiso wemfashini osezingeni eliphezulu kulabo abasembonini yezobuchwepheshe. Ngokuyibheka, ungathola kusenesikhathi ukuthi yimaphi amagama athakaselwa kakhulu kule sizini nokuthi yini ozoyizwa kuzo zonke izingqungquthela ezizayo.

Sithole ukuthi yini ebangela amagama amahle kule grafu ukuze ukwazi ukukhuluma nolimi.

Okokuqala, amagama ambalwa nje mayelana nokuthi hlobo luni lwegrafu lolu. Njalo ngonyaka ngo-Agasti, i-ejensi yokubonisana i-Gartner ikhipha umbiko - i-Gartner Hype Curve. NgesiRashiya, leli “ijika le-hype,” noma, kalula nje, i-hype. Eminyakeni engu-30 edlule, abaculi bomculo we-rap beqembu iPublic Enemy bacula: “Ungakukholelwa lokho okushiwo yilo mculo.” Ukholwa noma ungakholwa, umbuzo womuntu siqu, kodwa kufanelekile ukwazi okungenani lawa magama angukhiye uma usebenza emkhakheni wezobuchwepheshe futhi ufuna ukwazi izitayela zomhlaba.

Lena igrafu yalokho okulindelwe umphakathi okuvela kubuchwepheshe obuthile. NgokukaGartner, ngokufanelekile, ubuchwepheshe budlula ezigabeni ezi-5: ukwethulwa kobuchwepheshe, inani eliphakeme lokulindelwe, isigodi sokudumala, umthambeka wokukhanya, ithafa lokukhiqiza. Kodwa futhi kuyenzeka ukuthi iminza "esigodini sokudumazeka" - ungakhumbula izibonelo ngokwakho kalula, thatha ama-bitcoins afanayo: ekuqaleni ukushaya inani eliphakeme ngokuthi "imali yesikhathi esizayo", bashelela phansi ngokushesha lapho ukushiyeka kobuchwepheshe. kwacaca, okokuqala kukho konke imingcele enanini lokuthengiselana kanye nenani elikhulu likagesi elidingekayo ukuze kukhiqizwe ama-bitcoins (okuvele kuhlanganisa izinkinga zemvelo). Futhi-ke, akufanele sikhohlwe ukuthi ishadi likaGartner liyisibikezelo nje: lapha, isibonelo, ungafunda okuningiliziwe. indatshana, lapho izibikezelo ezingagcwaliseki eziphawuleka kakhulu zihlungwa khona.

Ngakho-ke, ake sidlule eshadini elisha le-Gartner. Ubuchwepheshe buhlukaniswe ngamaqembu ama-5 amakhulu anezihloko:

1. I-AI ethuthukisiwe nezibalo
2. I-Postclassical Compute kanye namaComms
3. Ukuzwa nokuhamba
4. Umuntu Ongathandwa
5. I-Digital Ecosystems

1. I-AI ethuthukisiwe nezibalo

Kule minyaka eyi-10 edlule sibone ihora elihle kakhulu lokufunda ngokujulile. Lawa manethiwekhi asebenza ngempela kububanzi bawo bemisebenzi. Ngo-2018, u-Yann LeCun, uGeoffrey Hinton kanye no-Yoshua Bengio bathola i-Turing Award ngokuthola kwabo - umklomelo ohlonishwa kakhulu, ofana noMklomelo KaNobel kwisayensi yekhompyutha. Ngakho, amathrendi amakhulu kule ndawo, aboniswa eshadini:

1.1. Dlulisa Ukufunda

Awuqeqesheli inethiwekhi ye-neural kusukela ekuqaleni, kodwa thatha esesiqeqeshiwe futhi uyinikeze umgomo ohlukile. Kwesinye isikhathi lokhu kudinga ukuqeqeshwa kabusha kwengxenye yenethiwekhi, kodwa hhayi yonke inethiwekhi, eshesha kakhulu. Isibonelo, ukuthatha inethiwekhi ye-neural eyenziwe ngomumo i-ResNet50, eqeqeshwe kudathasethi ye-ImageNet1000, uzothola i-algorithm engahlukanisa izinto eziningi ezahlukene esithombeni ezingeni elijule kakhulu (amakilasi angu-1000 asekelwe kuzici ezikhiqizwe izendlalelo ezingu-50 ze-neural. inethiwekhi). Kodwa awudingi ukuqeqesha yonke leyo nethiwekhi, okungathatha izinyanga.

В inkambo eku-inthanethi I-Samsung "Amanethiwekhi we-Neural nombono wekhompyutha", isibonelo, ekugcineni Umsebenzi we-Kaggle ngokuhlukaniswa kwamapuleti abe ahlanzekile futhi angcolile, indlela iyaboniswa ukuthi emizuzwini emi-5 ikunikeza inethiwekhi ejulile ye-neural ekwazi ukuhlukanisa amapuleti angcolile kwahlanzekile, akhiwe ngokuya ngesakhiwo esichazwe ngenhla. Inethiwekhi yasekuqaleni yayingazi ukuthi ayeyini amapuleti nhlobo, yafunda kuphela ukuhlukanisa izinyoni ezinjeni (bona ImageNet).

Source: inkambo eku-inthanethi Samsung "Neural networks and computer vision"

Ukuze uthole i-Transfer Learning udinga ukwazi ukuthi yiziphi izindlela ezisebenzayo nokuthi yiziphi izakhiwo eziyisisekelo esezilungile ezitholakalayo. Sekukonke, lokhu kusheshisa kakhulu ukuvela kwezicelo ezisebenzayo zokufunda ngomshini.

1.2. I-Generative Adversarial Networks (GAN)

Lokhu kungenxa yalezo zimo lapho kunzima kakhulu ngathi ukwenza umgomo wokufunda. Lapho umsebenzi usondela ekuphileni kwangempela, uqondakala nakakhulu kithi ("letha itafula eliseduze nombhede"), kodwa kunzima kakhulu ukuwenza njengomsebenzi wezobuchwepheshe. I-GAN iwumzamo nje wokusisindisa kule nkinga.

Kunamanethiwekhi amabili asebenza lapha: eyodwa i-generator (I-Generative), enye i- discriminator (I-Adversarial). Inethiwekhi eyodwa ifunda ukwenza umsebenzi owusizo (hlukanisa izithombe, ukubona imisindo, ukudweba opopayi). Futhi enye inethiwekhi ifunda ukufundisa leyo nethiwekhi: inezibonelo zangempela, futhi ifunda ukuthola ifomula eyinkimbinkimbi engaziwa ngaphambili yokuqhathanisa imikhiqizo yengxenye ekhiqizayo yenethiwekhi nezinto zomhlaba wangempela (isethi yokuqeqesha) ngokusekelwe ezicini ezijulile ezibaluleke ngempela. : inombolo yamehlo, ukusondela kwesitayela sika-Miyazaki, ukuphimisela kwesiNgisi okulungile.

Isibonelo somphumela wenethiwekhi yokukhiqiza izinhlamvu ze-anime. Umthombo

Kodwa, kunjalo, kunzima ukwakha izakhiwo lapho. Akwanele nje ukuphonsa ama-neurons, adinga ukulungiswa. Futhi kufanele ufunde amasonto. Ozakwethu esikhungweni se-Samsung Artificial Intelligence basebenza ngesihloko se-GAN; lona omunye wemibuzo yabo ebalulekile yocwaningo. Ngokwesibonelo, kanje ukuthuthukiswa: Ukusebenzisa amanethiwekhi akhiqizayo ukuze kuhlanganiswe izithombe ezingokoqobo zabantu abanesimo esiguquguqukayo - isibonelo, ukwakha igumbi lokulinganisa elibonakalayo, noma ukuhlanganiswa kobuso, okunganciphisa inani lolwazi oludinga ukugcinwa noma ukudluliselwa ukuze kuqinisekiswe ividiyo yekhwalithi ephezulu. ukuxhumana, ukusakaza noma ukuvikela idatha yomuntu siqu.

Umthombo

1.3. I-AI echazekayo

Kweminye imisebenzi eyivelakancane, intuthuko ekwakhiweni kwezakhiwo ezijulile ngokuzumayo ilethe amakhono amanethiwekhi ajulile we-neural eduze kwamakhono abantu. Manje impi iyaqhubeka yokwandisa ububanzi bemisebenzi enjalo. Isibonelo, isicoci se-robot vacuum singahlukanisa kalula ikati nenja emhlanganweni obhekene ngqo. Kodwa ezimweni eziningi zokuphila, ngeke akwazi ukuthola ikati elilele phakathi kwelineni noma ifenisha (noma kunjalo, njengathi, ezimweni eziningi ...).

Siyini isizathu sokuphumelela kwamanethiwekhi e-neural ajulile? Bakha ukumelwa kwenkinga okusekelwe olwazini “olubonakala ngeso lenyama” (amaphikseli ezithombe, izinguquko kuvolumu yomsindo...), kodwa ezicini ezitholwe ngemva kokucubungula kusengaphambili lolu lwazi ngezingqimba ezingamakhulu amaningana zenethiwekhi ye-neural. Ngeshwa, lobu budlelwano bungase bungabi nanjongo, bungahambisani, noma buthwale iminonjana yokungapheleli kusethi yedatha yoqobo. Isibonelo, kunomdlalo omncane wekhompiyutha mayelana nokuthi ukusetshenziswa ngokungacabangi kwe-AI ekuqasheni kungaholela kukho Ukusinda Kokufanelekile Okungcono Kakhulu.

Uhlelo lokufaka amathegi ezithombeni lubhale umuntu ophekayo njengowesifazane, nakuba umuntu osesithombeni eyindoda (Umthombo). yona uqaphele eVirginia Institute.

Ukuhlaziya ubudlelwano obuyinkimbinkimbi nobujulile esingeke sikwazi ukuzenzela thina, izindlela ezichazekayo ze-AI ziyadingeka. Bahlela izici zamanethiwekhi e-neural ajulile ukuze ngemva kokuqeqeshwa, sikwazi ukuhlaziya ukumelwa kwangaphakathi inethiwekhi esikufundile, kunokuba simane sincike esinqumweni sayo.

1.4. I-Edge Analytics / AI

Yonke into enegama elithi Edge isho lokhu okulandelayo: ukudlulisa ingxenye yama-algorithms isuka efwini/iseva iye kudivayisi yokugcina/ ileveli yesango. I-algorithm enjalo izosebenza ngokushesha futhi ngeke idinge ukuxhumana neseva emaphakathi ukuze isebenze. Uma ujwayelene nokuthathwa “kweklayenti elizacile,” khona-ke lapha senza leli klayenti libe lijiyile kancane.
Lokhu kungabaluleka ku-inthanethi Yezinto. Isibonelo, uma umshini ushise ngokweqile futhi udinga ukupholiswa, kunengqondo ukusayina lokhu ngokushesha, ezingeni lesitshalo, ngaphandle kokulinda idatha ukuthi iye emafini futhi ukusuka lapho kuya kumphathi we-shift. Noma esinye isibonelo: izimoto ezizishayelayo zingakwazi ukuzitholela isimo sethrafikhi ngokwazo, ngaphandle kokuxhumana neseva emaphakathi.

Umthombo

Noma esinye isibonelo sokuthi kungani lokhu kubalulekile ngokombono wezokuphepha: uma uthayipha imibhalo ocingweni lwakho, ikhumbula amagama ajwayelekile kuwe, ukuze ngokuhamba kwesikhathi ikhibhodi yefoni ikwazi ukukuyalela kalula ngawo - lokhu kubizwa ngokuthi i-predictive. okokufaka umbhalo. Ukuthumela yonke into oyibhalayo kukhibhodi yakho esikhungweni sedatha kwenye indawo kungaba ukwephula ubumfihlo bakho futhi kungaphephile nje. Ngakho-ke, ukuqeqeshwa kwekhibhodi kwenzeka kuphela ngaphakathi kwedivayisi yakho ngokwayo.

1.5. I-AI Platform njengesevisi (AI PaaS)

I-PaaS - I-Platform-as-a-Service iyimodeli yebhizinisi lapho sithola khona ukufinyelela kunkundla ehlanganisiwe, okuhlanganisa ukugcinwa kwayo kwedatha okusekelwe emafini kanye nezinqubo esezilungile. Ngale ndlela, singakwazi ukuzikhulula emisebenzini yengqalasizinda futhi sigxile ngokugcwele ekukhiqizeni okuthile okuwusizo. Isibonelo sezinkundla ze-PaaS zemisebenzi ye-AI: I-IBM Cloud, i-Microsoft Azure, i-Amazon Machine Learning, i-Google AI Platform.

1.6. I-Adaptive Machine Learning (i-Adaptive ML)

Kuthiwani uma sivumela ubuhlakani bokwenziwa buzivumelanise ... Uyabuza - okungukuthi, kanjani?.. Ingabe kakade ayikavumelani nomsebenzi? Inkinga yilena: siklama ngokucophelela inkinga ngayinye enjalo ngaphambi kokwakha i-algorithm yobuhlakani bokwenziwa ukuyixazulula. Bazokuphendula - kuvela ukuthi leli ketanga lingenziwa libe lula.

Ukufunda ngomshini okuvamile kusebenza ngomgomo we-open-loop: ulungisa idatha, uqhamuke nenethiwekhi ye-neural (noma yini), qeqesha, bese ubheka izinkomba ezimbalwa, futhi uma uthanda yonke into, ungathumela inethiwekhi ye-neural kuma-smartphones. ukuxazulula izinkinga zabasebenzisi . Kodwa ezinhlelweni zokusebenza lapho kukhona idatha eningi kanye nemvelo yayo ishintsha kancane kancane, ezinye izindlela ziyadingeka. Lezo zinhlelo, ezizivumelanisa nezizifundisa ngokwazo, zihlelwe zaba izihibe ezivaliwe, ezizifundelayo (iluphu evaliwe), futhi kufanele zisebenze kahle.

Izinhlelo zokusebenza - lokhu kungaba izibalo zokusakaza (Stream Analytics), ngesisekelo lapho osomabhizinisi abaningi benza izinqumo, noma ukuphathwa kokukhiqiza okuguquguqukayo. Esilinganisweni sezinhlelo zokusebenza zamanje futhi kunikezwe izingozi eziqondwa kangcono kubantu, amasu akha isixazululo sale nkinga wonke aqoqwa ngaphansi kwegama lesambulela le-Adaptive AI.

Umthombo

Uma ubheka lesi sithombe, kunzima ukuqeda umuzwa wokuthi ungadli isinkwa se-futurologists - mabafundise irobhothi ukuphefumula ...

I-Postclassical Compute kanye namaComms

2.1. Ukuxhumana kweselula kwesizukulwane sesihlanu (5G)

Lesi yisihloko esithakazelisa kangangokuba sikubhekisa kuwe ngokushesha isihloko. Hhayi-ke, nasi isifinyezo esifushane. I-5G, ngokwandisa imvamisa yokudluliswa kwedatha, izokwenza isivinini se-inthanethi sisheshe ngendlela engafanele. Kunzima kakhulu ukuthi amagagasi amafushane adlule izithiyo, ngakho-ke ukuklama kwamanethiwekhi kuzohluka ngokuphelele: izikhathi ezingu-500 iziteshi eziyisisekelo ezengeziwe ziyadingeka.

Kanye nesivinini, sizothola izenzakalo ezintsha: imidlalo yesikhathi sangempela eneqiniso elingathandwa kwabathelisi esikubona, ukwenza imisebenzi eyinkimbinkimbi (njengokuhlinzwa) ngokusebenzisa i-telepresence, ukuvimbela izingozi nezimo ezinzima emigwaqweni ngokuxhumana phakathi kwemishini. Ngokwenothi le-prosaic: I-inthanethi yeselula ekugcineni izoyeka ukwehla phakathi nemicimbi emikhulu, njengomdlalo enkundleni yezemidlalo.

Umthombo wesithombe - Reuters, Niantic

2.2. Inkumbulo Yesizukulwane Esilandelayo

Lapha sikhuluma ngesizukulwane sesihlanu se-RAM - DDR5. AbakwaSamsung bamemezele ukuthi imikhiqizo esekwe ku-DDR2019 izotholakala ekupheleni kuka-5. Kulindeleke ukuthi inkumbulo entsha izoshesha ngokuphindwe kabili futhi ibe namandla ngokuphindwe kabili ngenkathi sigcina uhlobo olufanayo lwefomu, okungukuthi, sizokwazi ukuthola izinti zememori ezinomthamo ongafika ku-32GB wekhompyutha yethu. Ngokuzayo, lokhu kuzosebenza ikakhulukazi kuma-smartphones (inkumbulo entsha izoba kunguqulo yamandla aphansi) kanye namalaptop (lapho inani lezikhala ze-DIMM lilinganiselwe). Futhi ukufunda ngomshini nakho kudinga amanani amakhulu e-RAM.

2.3. I-Low-Earth-Orbit Satellite Systems

Umqondo wokushintsha amasathelayithi asindayo, abizayo, anamandla ngoxhaxha lwamancane futhi ashibhile usekude nokusha futhi wavela emuva kuma-90s. Mayelana nani "U-Elon Musk maduze uzosabalalisa i-inthanethi kuwo wonke umuntu osebenzisa isathelayithi" Manje amavila kuphela awazwanga. Inkampani edume kakhulu lapha i-Iridium, eyaqothuka ngasekupheleni kweminyaka yama-90s, kodwa yasindiswa ngezindleko zoMnyango Wezokuvikela wase-US (akumele kudidaniswe ne-iRidium, uhlelo lwasekhaya oluhlakaniphile lwaseRussia). Iphrojekthi ka-Elon Musk (Starlink) ikude nokuphela kwayo - uRichard Branson (ama-satellite ahlongozwayo e-OneWeb - 1440), i-Boeing (ama-satellites angu-3000), i-Samsung (ama-satellites angu-4600), nabanye babambe iqhaza emncintiswaneni we-satellite.

Ukuthi izinto zimi kanjani kule ndawo, ukuthi umnotho ubukeka kanjani lapho - funda ngaphakathi ukubuyekeza. Futhi silindele ukuhlolwa kokuqala kwalezi zinhlelo ngabasebenzisi bokuqala, okufanele kwenzeke ngonyaka ozayo.

2.4. I-Nanoscale 3D Printing

Ukuphrinta kwe-3D, nakuba kungakangeni empilweni yawo wonke umuntu (ngendlela ethenjiswe ngayo imboni yepulasitiki yasekhaya), nokho kudala kwashiya i-niche yobuchwepheshe yama-geek. Ungahlulela ngokuthi yonke ingane yesikole iyazi mayelana nokuba khona okungenani kwamapeni aqoshiwe e-3D, futhi abaningi baphupha ngokuthenga ibhokisi elinabagijimi kanye ne-extruder ye... “nje kanjalo” (noma sebevele balithengile).

I-Stereolithography (amaphrinta e-laser 3D) ivumela ukuphrinta ngama-photon angawodwana: ama-polymer amasha ayahlolwa adinga kuphela ama-photon amabili ukuze aqine. Lokhu kuzovumela, ezimweni ezingezona ezebhu, ukudala izihlungi ezintsha ngokuphelele, izintaba, iziphethu, ama-capillaries, ama-lens kanye ... izinketho zakho kumazwana! Futhi lapha akukude ne-photopolymerization - lobu buchwepheshe kuphela buvumela ukuthi "siphrinte" ama-processor kanye namasekethe ekhompiyutha. Ngaphezu kwalokho, lona akuwona unyaka wokuqala ukuba khona ubuchwepheshe bokuphrinta i-graphene 500 nm izakhiwo ezintathu-ntathu, kodwa ngaphandle kwentuthuko enkulu.

Umthombo

3. Ukuzwa Nokuhamba

3.1. I-Autonomous Driving Level 4 & 5

Ukuze ungadideki kumagama, kufanelekile ukuqonda ukuthi yimaphi amazinga okuzimela ahlukaniswa (athathwe emininingwaneni izindatshana, lapho sibhekisa khona bonke abanentshisekelo):

Izinga 1: I-Cruise control: isiza umshayeli ezimeni ezilinganiselwe kakhulu (isibonelo, ukubamba imoto ngesivinini esinikeziwe ngemva kokuba umshayeli esuse unyawo lwakhe ku-pedal)
Izinga 2: Usizo olulinganiselwe lokuqondisa nokubhuleka. Umshayeli kufanele alungele ukulawula cishe ngaso leso sikhathi. Izandla zisesiteringini, amehlo abheke emgwaqeni. Lokhu yinto uTesla noGeneral Motors asebenayo.
Izinga 3: Umshayeli akusadingeki ukuthi ahlale ebuka umgwaqo. Kodwa kufanele ahlale eqaphile futhi alungele ukulawula. Lena into izimoto ezithengiswayo ezingakabi nayo. Zonke ezikhona njengamanje zikuleveli 1-2.
Izinga lesi-4: Ukushayela okuzenzakalelayo kweqiniso, kodwa okunemikhawulo: uhambo kuphela endaweni eyaziwayo edwetshwe ngokucophelela futhi eyaziwa ngokuvamile ohlelweni, nangaphansi kwezimo ezithile: isibonelo, lapho lingekho iqhwa. I-Waymo ne-General Motors banama-prototypes anjalo, futhi bahlela ukuyethula emadolobheni amaningana futhi bawahlole ezindaweni zangempela. I-Yandex inezindawo zokuhlola amatekisi angenamuntu e-Skolkovo nase-Innopolis: uhambo lwenzeka ngaphansi kokuqondisa kukanjiniyela ohlezi esihlalweni somgibeli; ekupheleni konyaka, inkampani ihlela ukwandisa imikhumbi yayo ibe yizimoto ezinganamuntu eziyi-100.
Izinga lesi-5: Ukushayela okuzenzakalelayo okugcwele, ukushintshwa okuphelele komshayeli ophilayo. Izinhlelo ezinjalo azikho, futhi cishe ngeke zivele eminyakeni ezayo.

Kunengqondo kangakanani ukubona konke lokhu esikhathini esizayo? Lapha ngithanda ukuqondisa kabusha umfundi esihlokweni "Kungani kungenakwenzeka ukwethula irobhothi ngo-2020, njengoba uTesla ethembisa". Lokhu ngokwengxenye kungenxa yokuntuleka kokuxhumeka kwe-5G: isivinini esitholakalayo se-4G azanele. Ngokwengxenye ngenxa yezindleko eziphakeme kakhulu zezimoto ezizimele: azikabi nenzuzo, imodeli yebhizinisi ayicacile. Ngamafuphi nje, “konke kuyinkimbinkimbi” lapha, futhi akukhona nje ukuthi u-Gartner abhale ukuthi isibikezelo sokuqaliswa kobuningi beZinga lesi-4 nelesi-5 asikabi ngaphambi kweminyaka eyi-10.

3.2. Amakhamera Enzwa we-3D

Eminyakeni eyisishiyagalombili edlule, isilawuli segeyimu se-Kinect seMicrosoft senza amagagasi ngokunikeza isixazululo esifinyelelekayo nesishibhile kumbono we-3D. Kusukela ngaleso sikhathi, imidlalo yemfundo engokomzimba nemidlalo yokudansa ne-Kinect ihlangabezane nokukhuphuka nokuncipha kwayo okufushane, kodwa amakhamera e-3D aqala ukusetshenziswa kumarobhothi ezimboni, ezimotweni ezingenamuntu, kanye nomakhalekhukhwini ukuze kuhlonzwe ubuso. Ubuchwepheshe sebushibhile, buhlangene futhi bufinyeleleka kalula.

Ifoni ye-Samsung S10 inekhamera ye-Time-of-Flight ekala ibanga ukuya entweni ukwenza ukugxila kube lula. Umthombo

Uma unentshisekelo kulesi sihloko, sizobe sesikuqondisa kabusha ekubuyekezweni okuhle kakhulu okunemininingwane yamakhamera ajulile: ingxenye 1, ingxenye 2.

3.3. Ama-Drones okuletha impahla encane (IziDrone Zokulethwa Kwezimpahla Ezikhanyayo)

Kulo nyaka, i-Amazon yenze amagagasi lapho ikhombisa ngendiza ezindizayo entsha embukisweni ekwazi ukuthwala imithwalo emincane efika ku-2 kg. Edolobheni elinesiminyaminya, lokhu kubonakala kuyisixazululo esifanelekile. Ake sibone ukuthi la ma-drones asebenza kanjani esikhathini esizayo esiseduze. Mhlawumbe kufanelekile ukungabaza ngokucophelela lapha: kunezinkinga eziningi, eziqala ngokuthi kungenzeka ukwebiwa kalula kwe-drone, futhi kugcine ngemikhawulo engokomthetho kuma-UAV. I-Amazon Prime Air isineminyaka eyisithupha ikhona kodwa isesigabeni sokuhlola.

I-drone entsha yase-Amazon, ekhonjiswe kule ntwasahlobo. Kukhona i-Star Wars ngaye. Umthombo

Ngaphezu kwe-Amazon, kukhona abanye abadlali kule makethe (kunemininingwane eningiliziwe ukubuyekeza), kodwa akuwona umkhiqizo owodwa oqediwe: yonke into isesigabeni sokuhlola nemikhankaso yokumaketha. Ngokwehlukana, kufanelekile ukuphawula ezithakazelisayo kakhulu zezokwelapha ezikhethekile amaphrojekthi e-Afrika: ukulethwa kwegazi elinikelwe eGhana (ukulethwa kwe-14, inkampani ye-Zipline) naseRwanda (inkampani ye-Matternet).

3.4. Izimoto Ezizihambelayo Ezindizayo

Kunzima ukusho noma yini eqinisekile lapha. Ngokusho kukaGartner, lokhu ngeke kuvele ngaphambi kweminyaka eyi-10. Ngokuvamile, kukhona zonke izinkinga ezifanayo lapha ezimotweni ezizishayelayo, kuphela zithola isici esisha - mpo. IPorsche, Boeing, kanye no-Uber bamemezele izifiso zabo zokwakha itekisi endizayo.

3.5. I-Augmented Reality Cloud (Ifu Le-AR)

Ikhophi ehlala njalo yedijithali yomhlaba wangempela, ekuvumela ukuthi udale isendlalelo esisha sangempela esivamile kubo bonke abasebenzisi. Ngamanye amagama obuchwepheshe, sikhuluma ngokwenza inkundla yefu evulekile lapho onjiniyela bengahlanganisa khona izinhlelo zabo zokusebenza ze-AR. Imodeli yokwenza imali icacile; iwuhlobo lwe-analogue ye-Steam. Umbono usugxile kangangokuthi abanye manje bakholelwa ukuthi i-AR ngaphandle kwefu imane nje iyize.

Ukuthi lokhu okungase kubukeke kanjani esikhathini esizayo kuboniswa kuvidiyo emfushane. Kubukeka njengesinye isiqephu se-Black Mirror:

Ungafunda futhi ku- buyekeza isihloko.

4. Ukungathandwa Kwabantu

4.1. I-Emotion AI

Indlela yokulinganisa, ukulingisa nokusabela emizweni yabantu? Amanye amaklayenti lapha yizinkampani ezenza abasizi bezwi njenge-Amazon Alexa. Bangajwayela ngempela amakhaya uma befunda ukubona isimo: baqonde isizathu sokunganeliseki komsebenzisi, futhi bazame ukulungisa isimo. Ngokuvamile, kunolwazi oluningi kumongo kunasemyalezo ngokwawo. Futhi umongo isimo sobuso, iphimbo, nokuziphatha okungakhulumi.

Ezinye izinhlelo zokusebenza ezingokoqobo: ukuhlaziya imizwa phakathi nenhlolokhono yomsebenzi (okusekelwe ezingxoxweni zevidiyo), ukuhlola ukusabela ezimakethe noma okunye okuqukethwe kwamavidiyo (ukumomotheka, ukuhleka), usizo lokufunda (ngokwesibonelo, ukuzijwayeza okuzimele kubuciko bokukhuluma esidlangalaleni).

Kunzima ukukhuluma kangcono ngalesi sihloko kunombhali wefilimu emfushane yemizuzu engu-6 Ukweba i-Ur Feeling. Ividiyo enobuhlakani nesitayela ibonisa ukuthi ungayikala kanjani imizwa yethu ngezinjongo zokumaketha, futhi kusukela ekuphenduleni kwesikhashana kobuso bakho, thola ukuthi ingabe uyayithanda i-pizza, izinja, i-Kanye West, kanye nokuthi liyini izinga lakho lemali engenayo kanye ne-IQ elinganiselwe. Ngokuvakashela iwebhusayithi yefilimu usebenzisa isixhumanisi esingenhla, uba umhlanganyeli kuvidiyo esebenzisanayo usebenzisa ikhamera eyakhelwe ngaphakathi yekhompyutha yakho ephathekayo. Le filimu isivele ikhonjisiwe emicimbini eminingi yamafilimu.

Umthombo

Kukhona nesifundo esithakazelisa kangako: indlela yokubona ukubhuqa embhalweni. Sithathe ama-tweets ane-hashtag ethi #sarcasm futhi senza isethi yokuqeqesha yama-tweets angu-25 anenhlamba kanye nama-tweets angu-000 avamile mayelana nayo yonke into ngaphansi kwelanga. Sisebenzise umtapo wezincwadi we-TensorFlow, saqeqesha uhlelo, futhi nawu umphumela:

Umthombo

Ngakho-ke, manje, uma ungaqiniseki ngozakwenu noma umngane wakho - usho okuthile kuwe ngokungathí sina noma ngokubhuqa, usungakwazi kakade ukusebenzisa inethiwekhi ye-neural eqeqeshiwe!

4.2. I-Augmented Intelligence

Ukuzenzakalela komsebenzi wobuhlakani kusetshenziswa izindlela zokufunda ngomshini. Kungase kubonakale sengathi akukho okusha? Kodwa amagama ngokwawo abalulekile lapha, ikakhulukazi njengoba ehambisana nesifinyezo ne-Artificial Intelligence. Lokhu kusibuyisela engxoxweni mayelana ne-AI “eqinile” kanye “nobuthakathaka”.
I-AI eqinile iwubuhlakani bokwenziwa obufanayo obuvela kumafilimu ezinganekwane zesayensi obulingana ngokuphelele nomqondo womuntu futhi ozaziyo njengomuntu ngamunye. Lokhu akukabi khona futhi akucaci noma kuzoba khona nhlobo.

I-AI ebuthakathaka ayiyena umuntu ozimele, kodwa umsizi womuntu. Akasho ukuthi unokucabanga okomuntu, kodwa umane wazi indlela yokuxazulula izinkinga zolwazi, isibonelo, ukunquma ukuthi yini eboniswa esithombeni noma ukuhumusha umbhalo.

Umthombo

Ngalo mqondo, i-Augmented Intelligence "i-AI ebuthakathaka" ngendlela yayo ehlanzekile, futhi ukwakheka kubonakala kuphumelele, ngoba akwethuli ukudideka nesilingo sokubona lapha "i-AI eqinile" wonke umuntu ayiphuphayo (noma ukwesaba, uma khumbula izingxoxo eziningi mayelana "nezimoto zokuvukela"). Sisebenzisa inkulumo ethi Augmented Intelligence, ngokushesha siba amaqhawe enye ifilimu: kusukela ezinganekwaneni zesayensi (njengeka-Asimov ethi “I, Robot”) sizithola siku-cyberpunk (“ama-augmentations” kulolu hlobo yizo zonke izinhlobo zokufakelwa okwandisa amandla omuntu).

Indlela kusho U-Erik Brynjolfsson no-Andrew McAffee: “Eminyakeni eyi-10 ezayo, yilokhu okuzokwenzeka. I-AI ngeke ingene esikhundleni sabaphathi, kodwa labo baphathi abasebenzisa i-AI bazongena esikhundleni salabo abangakayingeni.”

izibonelo:

• Umuthi: IYunivesithi yaseStanford yathuthukiswa i-algorithm, obhekana nomsebenzi wokuqaphela izifo ze-X-ray esifubeni ngokwesilinganiso ngempumelelo njengodokotela abaningi
• Imfundo: usizo kubafundi nothisha, ukuhlaziya izimpendulo zabafundi ezintweni zokwakha, ukwakhiwa kwendlela yokufunda yomuntu ngamunye.
• Izibalo zebhizinisi: ukucubungula idatha, ngokwezibalo, kuthatha u-80% wesikhathi somcwaningi, futhi kuphela u-20% wokuhlolwa ngokwako.

4.3. Ama-Biochips

Lena itimu eyintandokazi yawo wonke amafilimu nezincwadi ze-cyberpunk. Ngokuvamile, izilwane ezifuywayo ezincane akuwona umkhuba omusha. Kodwa manje la ma-chips aseqalile ukufakwa kubantu.

Kulokhu, i-hype cishe ihlotshaniswa necala elivusa amadlingozi enkampanini yaseMelika i-Three Square Market. Lapho, umqashi waqala ukunikela ngokufaka ama-chips ngaphansi kwesikhumba ukuze akhokhe imali. I-chip ikuvumela ukuthi uvule iminyango, ungene kumakhompiyutha, uthenge ukudla okulula emshinini othengisayo - okungukuthi, ikhadi elinjalo lomsebenzi wonke. Ngaphezu kwalokho, i-chip enjalo isebenza njengekhadi lokuhlonza; ayinayo imojula ye-GPS, ngakho-ke akunakwenzeka ukulandelela noma ubani oyisebenzisayo. Futhi uma umuntu efuna ukususa i-chip engalweni yakhe, kuthatha imizuzu emi-5 ngosizo lukadokotela.

Ama-chips ngokuvamile afakwa phakathi kwesithupha nomunwe wokukhomba. Umthombo

Funda kabanzi indatshana mayelana nesimo sezindaba ngokuchichima emhlabeni.

4.4. Indawo Yokusebenzela Egxilile

“Ukucwila” elinye igama elisha elingenakho ukubalekela. Ikhona yonke indawo. Itiyetha egxilile, umbukiso, icinema. Usho ukuthini? Ukucwiliswa ukudalwa komphumela wokucwiliswa, lapho umngcele phakathi kombhali nesibukeli, okubonakalayo nomhlaba wangempela kulahleka. Ezindaweni zokusebenza, cishe, lokhu kusho ukufiphalisa umugqa phakathi komenzi kanye nomqalisi kanye nokukhuthaza abasebenzi ukuthi bathathe indawo esebenzayo ngokufometha kabusha indawo yabo.

Njengoba manje sine-Agile, evumelana nezimo, nokubambisana eduze yonke indawo, izindawo zokusebenza kufanele zilungiseke kalula ngangokunokwenzeka futhi kufanele zikhuthaze umsebenzi weqembu. Umnotho ubeka imigomo yawo: kunezisebenzi zesikhashana ezengeziwe, izindleko zokuqasha indawo yamahhovisi ziyenyuka, futhi emakethe yezabasebenzi enokuncintisana, izinkampani ze-IT zizama ukukhulisa ukwaneliseka kwabasebenzi emsebenzini ngokudala izindawo zokungcebeleka nezinye izinzuzo. Futhi konke lokhu kubonakala ekwakhiweni kwezindawo zokusebenza.

Из umbiko UKnoll

4.5. Ukwenza umuntu

Wonke umuntu uyazi ukuthi kuyini ukwenza kube ngokwakho ekukhangiseni. Kulapho namuhla uxoxa nozakwenu ukuthi umoya osegumbini womile, futhi kufanele uthenge isihwamulisa-manzi sehhovisi, futhi ngakusasa ubona isikhangiso kunethiwekhi yakho yokuxhumana nabantu - “thenga okokuhudisa umoya” (a isigameko sangempela esenzeke kimi).

Umthombo

Ukwenza kube ngokwakho, njengoba kuchazwe u-Gartner, kuyimpendulo yokukhathazeka okukhulayo kwabasebenzisi mayelana nokusetshenziswa kwedatha yabo yomuntu siqu ngezinjongo zokukhangisa. Umgomo uwukuthuthukisa indlela esiboniswa ngayo ukukhangisa ehambisana nengqikithi esizithola sikuyo, hhayi kithina mathupha. Isibonelo, indawo yethu, uhlobo lwedivayisi, isikhathi sosuku, izimo zezulu - lokhu kuyinto engaphuli idatha yethu yomuntu siqu, futhi asizwa umuzwa ongemuhle "wokubhekwa".

Funda ngomehluko phakathi kwale miqondo emibili inothi U-Andrew Frank ubhalisa kuwebhusayithi kaGartner. Kukhona umehluko ocashile namagama afanayo kangangokuthi wena, ungazi umehluko, ubeka engozini ukuphikisana nomuntu wakho isikhathi eside, ungasoli ukuthi, ngokuvamile, bobabili baqinisile (futhi lokhu kuyisigameko sangempela esenzeka ku-interlocutor yakho. umbhali).

4.6. I-Biotech - Izicubu Ezakhiwe noma Ezokwenziwa

Lokhu, okokuqala, umqondo wokukhulisa inyama yokwenziwa. Ngesikhathi esifanayo, amaqembu amaningana emhlabeni wonke amatasa ukuthuthukisa ilabhorethri "Inyama 2.0" - kulindeleke ukuthi izoba eshibhile kunokuvamile, futhi ukudla okusheshayo kanye nezitolo ezinkulu zizoshintshela kuyo. Abatshalizimali kulobu buchwepheshe bahlanganisa uBill Gates, uSergey Brin, uRichard Branson nabanye.

Umthombo

Izizathu zokuthi kungani wonke umuntu enentshisekelo kangaka ngenyama yokwenziwa:

1. Ukufudumala kwembulunga yonke: ukukhishwa kwe-methane okuvela emapulazini. Lokhu kungama-18% omthamo wamagesi omhlaba wonke athinta isimo sezulu.
2. Ukwanda kwabantu. Isidingo senyama siyakhula, futhi ngeke kwenzeke ukondla wonke umuntu ngenyama yemvelo - kuyabiza nje.
3. Ukuntuleka kwendawo. Amaphesenti angama-70 amahlathi ase-Amazon asegawuliwe ukuze adliwe.
4. Ukucatshangelwa kokuziphatha. Kukhona labo okubalulekile kubo lokhu. Inhlangano elwela amalungelo ezilwane i-PETA isivele inikeze umklomelo ka-$1 million kusosayensi oletha inyama yenkukhu yokwenziwa emakethe.

Ukushintsha inyama yangempela ngesoya kuyisixazululo esiyingxenye, ngoba abantu bangawuqonda umehluko ekunambithekeni nasekuthunjweni, futhi akunakwenzeka ukuthi bayeke i-steak ngokuthanda isoya. Ngakho-ke udinga inyama yangempela, etshalwe ngokwemvelo. Manje, ngeshwa, inyama yokwenziwa ibiza kakhulu: kusuka ku-$ 12 ngekhilogremu. Lokhu kungenxa yenqubo yezobuchwepheshe eyinkimbinkimbi yokukhulisa inyama enjalo. Funda ngakho konke indatshana.

Uma sikhuluma ngezinye izimo zokukhula kwezicubu - kakade kwezokwelapha - isihloko esinezitho zokwenziwa siyathakazelisa: isibonelo, "isiqephu" senhliziyo, ephrintiwe iphrinta ye-3D ekhethekile. Kwaziwa izindaba njengenhliziyo yegundane elikhule ngokuzenzela, kodwa ngokuvamile yonke into isangaphakathi kobubanzi bokuhlolwa komtholampilo. Ngakho-ke mancane amathuba okuba sibone iFrankenstein eminyakeni ezayo.

Lapha u-Gartner uqaphe kakhulu ezilinganisweni zakhe, ngokusobala ekhumbula isibikezelo sakhe esingaphumelelanga sika-2015 sokuthi ngo-2019, u-10% wabantu emazweni athuthukile uzoba nokufakelwa kwemishini yezokwelapha ephrintiwe eyi-3D. Ngakho-ke, kusho ukuthi isikhathi sokufinyelela endaweni yokukhiqiza okungenani iminyaka eyi-10.

5. I-Digital Ecosystems

5.1. Iwebhu Eyehlukaniswe

Lo mqondo uhlobene eduze negama lomsunguli wewebhu, owine i-Turing Award uSir Tim Burners-Lee. Kuye, imibuzo yezimiso zokuziphatha kwisayensi yekhompiyutha yayihlale ibalulekile futhi ingqikithi ehlangene ye-intanethi yayibalulekile: ukubeka izisekelo ze-hypertext, wayeqiniseka ukuthi inethiwekhi kufanele isebenze njengewebhu, hhayi njenge-hierarchy. Lokhu kwakunjalo ekuqaleni kokuthuthukiswa kwenethiwekhi. Nokho, njengoba i-inthanethi ikhula, ukwakheka kwayo kwaba phakathi kwezizathu ezihlukahlukene. Kuvele ukuthi ukufinyelela kunethiwekhi yezwe lonke kungavinjwa kalula ngosizo lwabahlinzeki abambalwa. Futhi idatha yomsebenzisi isiphenduke umthombo wamandla nemali engenayo ezinkampanini ze-inthanethi.

UBurners-Lee uthi: “I-Internet isivele yasakazwa ezindaweni eziningi. “Inkinga ukuthi injini yokusesha eyodwa, inkundla yokuxhumana enkulukazi, inkundla eyodwa ye-microblogging ibusa. Asinazo izinkinga zobuchwepheshe, kodwa sinazo ezenhlalo.”

Kwakhe incwadi evuliwe Ngeminyaka engama-30 ye-World Wide Web, umdali weWebhu waveza izinkinga ezintathu eziyinhloko ze-inthanethi:

1. Ukulimala okuhlosiwe njengokugebenga okuxhaswe nguhulumeni, ubugebengu, nokuhlukumeza ku-inthanethi
2. Yona kanye idizayini yesistimu, elimaza umsebenzisi, idala isisekelo sezindlela ezinjengalezi: izikhuthazo zezimali ze-clickbait kanye nokusabalala kolwazi olungamanga.
3. Imiphumela engahlosiwe yedizayini yesistimu eholela ekungqubuzaneni nekhwalithi encishisiwe yengxoxo ye-inthanethi

Futhi u-Tim Berners-Lee usenayo impendulo yokuthi yiziphi izimiso "I-inthanethi Yomuntu Onempilo" engase isekelwe kuyo, engenayo inombolo yenkinga yesi-2: "Kubasebenzisi abaningi, imali engenayo yokukhangisa ihlala iwukuphela kwemodeli yokuxhumana ne-intanethi. Ngisho noma abantu besaba ngokwenzeka kudatha yabo, bazimisele ukwenza isivumelwano nomshini wokumaketha ukuze bathole ithuba lokuthola okuqukethwe mahhala. Cabanga ngezwe lapho ukukhokhela izimpahla nezinkonzo kulula futhi kujabulisa kuzo zombili izinhlangothi.” Phakathi kwezinketho zokuthi lokhu kungahlelwa kanjani: abaculi bangathengisa okurekhodiwe kwabo ngaphandle kwabaxhumanisi ngendlela ye-iTunes, futhi izingosi zezindaba zingasebenzisa uhlelo lwe-micropayments ukufunda isihloko esisodwa, esikhundleni sokwenza imali ngokukhangisa.

Njengesibonelo sokuhlola sale Inthanethi entsha, u-Tim Berners-Lee wethule iphrojekthi ye-SOLID, ingqikithi yayo ukuthi ugcina idatha yakho “ku-pod” - isitolo solwazi, futhi anganikeza le datha ezinhlelweni zezinkampani zangaphandle. Kodwa empeleni, wena ngokwakho ungabaphathi bedatha yakho. Konke lokhu kuhlobene eduze nomqondo wamanethiwekhi ontanga kuya kontanga, okungukuthi, ikhompyutha yakho ayiceli kuphela izinsizakalo, kodwa futhi inikeza, ukuze ungathembeli kuseva eyodwa njengesiteshi kuphela.

Umthombo

5.2. Izinhlangano Ezizimele Ezizimele

Yinhlangano elawulwa yimithetho ebhalwe phansi ngohlelo lwe-computer. Imisebenzi yayo yezezimali isekelwe ku-blockchain. Inhloso yokudala izinhlangano ezinjalo ukuqeda umbuso endimeni yokuba umlamuli nokudala indawo ethembekile yokwethenjwa kozakwethu, okungesiyo okwanoma ubani umuntu ngamunye, kodwa okuphethwe yiwo wonke umuntu ngokuhlanganyela. Okusho ukuthi, ngokombono, lokhu kufanele, uma umbono ugxila, kuqedwe ama-notaries nezinye izikhungo zokuqinisekisa ezijwayelekile.

Isibonelo esidume kakhulu senhlangano enjalo kwakuyi-DAO egxile ebhizinisini, eyaqoqa amaRandi ayizigidi ezingu-2016 ngo-150, lapho i-$ 50 yantshontshwa ngokushesha ngembobo esemthethweni emithethweni. Kwavela inkinga enzima ngokushesha: noma uhlehlise futhi ubuyisele imali, noma uvume ukuthi ukuhoxiswa kwemali kwakungokomthetho, ngoba akuzange kuphule imithetho yesikhulumi. Ngenxa yalokho, ukuze kubuyiselwe imali kubatshalizimali, abadali kwakudingeka babhubhise i-DAO, babhale kabusha i-blockchain futhi baphule isimiso sayo esiyisisekelo - ukungaguquki.

Amahlaya mayelana ne-Ethereum (kwesokunxele) kanye ne-DAO (kwesokudla). Umthombo

Yonke le ndaba yonakalise isithunzi sawo kanye umqondo we-DAO. Leyo phrojekthi yenziwa ngesisekelo se-cryptocurrency ye-Ethereum, inguqulo ka-Etere 2.0 kulindeleke ngonyaka ozayo - mhlawumbe abalobi (phakathi kwabo odumile Vitalik Buterin) bazocabangela amaphutha futhi babonise into entsha. Kungenzeka ukuthi yingakho u-Gartner ebeka i-DAO ku-upline.

5.3.Idatha Yezokwenziwa

Ukuze uqeqeshe amanethiwekhi e-neural, inani elikhulu ledatha liyadingeka. Ukulebula idatha mathupha kuwumsebenzi omkhulu ongenziwa umuntu kuphela. Ngakho-ke, kungenzeka ukudala amasethi edatha yokwenziwa. Isibonelo, amaqoqo afanayo obuso bomuntu kusayithi https://generated.photos. Zakhiwe kusetshenziswa i-GAN - ama-algorithms aseshiwo ngenhla.

Lobu buso akubona obabantu. Umthombo

Inzuzo enkulu yedatha enjalo ukuthi abukho ubunzima obungokomthetho bokuyisebenzisa: akekho onikeza imvume yokucutshungulwa kwedatha yomuntu siqu.

5.4.Ama-Digital Ops

Isijobelelo esithi "Ops" sesiphenduke imfashini ngendlela emangalisayo kusukela kwaqala i-DevOps enkulumweni yethu. Manje mayelana nokuthi iyini i-DigitalOps - iwukuhlanganisa nje i-DevOps, i-DesignOps, i-MarketingOps... Ingabe usunesithukuthezi okwamanje? Ngamafuphi, ukudluliselwa kwendlela ye-DevOps isuka endaweni yesoftware iye kuzo zonke ezinye izici zebhizinisi - ukumaketha, ukuklama, njll.

Umthombo

Umqondo we-DevOps bekuwukususa izithiyo phakathi kweNtuthuko ngokwayo kanye Nokusebenza (izinqubo zebhizinisi), ngokwakhiwa kwamaqembu ajwayelekile, lapho kukhona abahleli bezinhlelo, abahloli, ochwepheshe bezokuphepha, nabaphathi; ukuqaliswa kwezinqubo ezithile: ukuhlanganiswa okuqhubekayo, ingqalasizinda njengekhodi, ukuncishiswa nokuqinisa iketango lempendulo. Inhloso bekuwukusheshisa isikhathi sokuthengisa somkhiqizo. Uma ucabanga ukuthi lokhu kufana ne-Agile, uqinisile. Manje dlulisela ngokwengqondo le ndlela kusuka emkhakheni wokuthuthukiswa kwesoftware uye ekuthuthukisweni ngokujwayelekile - futhi uyaqonda ukuthi iyini i-DigitalOps.

5.5. Amagrafu olwazi

Indlela yesofthiwe yokumodela indawo yolwazi, okuhlanganisa ukusebenzisa ama-algorithms okufunda komshini. Igrafu yolwazi yakhiwe phezu kolwazi olukhona ukuze kuxhunywe ndawonye lonke ulwazi: kokubili okuhlelekile (uhlu lwezehlakalo noma abantu) nokungahlelekile (umbhalo we-athikili).

Isibonelo esilula yikhadi ongalibona emiphumeleni yosesho lwe-Google. Uma ufuna umuntu noma isikhungo, uzobona ikhadi ngakwesokudla:


Sicela uqaphele ukuthi "Imicimbi Ezayo" akuyona ikhophi yolwazi oluvela ku-Google Amamephu, kodwa ukuhlanganiswa kweshejuli ne-Yandex.Afisha: ungabona lokhu kalula uma uchofoza ezenzakalweni. Okusho ukuthi, inhlanganisela yemithombo yedatha eminingana ndawonye.

Uma ucela uhlu - isibonelo, "abaqondisi abadumile" - uzoboniswa i-carousel:

Ibhonasi kulabo abafunde kuze kube sekugcineni

Futhi manje njengoba sesizicacisele ngokwethu incazelo yephuzu ngalinye, singabheka isithombe esifanayo, kodwa ngesiRashiya:

Yabelana ngokukhululekile ezinkundleni zokuxhumana!

U-Tatyana Volkova - Umbhali wohlelo lokuqeqeshwa lwe-inthanethi Yezinto IT track e-Samsung Academy, uchwepheshe wezinhlelo zezinkampani zokubophezeleka komphakathi e-Samsung Research Center.


Source: www.habr.com