Ingxoxo yesikhashana ye-ChatGPT isibe "imodi ye-incognito" ye-AI.Lesi sici senzelwe labo abafuna ukubuza imibuzo ebucayi, ukuhlola imibono eyimfihlo, noma ukuvimba ingxoxo ethile ukuthi ingathinti ezinye izingxoxo zabo. Kodwa lapho sisetshenziswa kakhulu, kulapho kuvela khona ukungabaza okwengeziwe: Siyimfihlo kangakanani ngempela? Yini egcinwa yi-OpenAI? Ingabe uhlelo "lungakhumbula" into okuthiwa aluyigcini?
Ezinyangeni zakamuva, ubufakazi obuvela kubasebenzisi abakhathazekile buvele. abathi babone ulwazi olufakwe engxoxweni yesikhashana luvela kabusha engxoxweni evamile, ngaphezu kwezingxoxo zobuchwepheshe mayelana nendlela lolu lwazi "oluyisikhashana" oluphathwa ngayo, njenge Ukwehluleka komshayeli wendizaNgesikhathi esifanayo, imibhalo ye-OpenAI ichaza ukuthi ikhophi igcinwa izinsuku ezingu-30 ngenxa yezizathu zokuphepha. Konke lokhu kudale ingxubevange yesithakazelo, ukudideka, kanye nokungathembani okufanele kucaciswe ngokuthula.
Iyini ngempela ingxoxo yesikhashana ye-ChatGPT?
Ingxoxo yesikhashana ye-ChatGPT iyindlela yokuxoxa "ekhethekile" eyenzelwe ukwandisa ubumfihlo.Ukuyisebenzisa kuqala iseshini engagcinwanga emlandweni wakho wokuphequlula ojwayelekile futhi, ngokusho kwe-OpenAI, ayisetshenziswa ukuqeqesha amamodeli ayo e-AI. Empeleni, ifana nemodi ye-incognito kuziphequluli ezifana ne-Chrome, i-Safari, noma i-Firefox, kodwa isetshenziswa kumsizi wengxoxo.
Engxoxweni evamile, zonke izingxoxo zakho ziqoshwa ku-akhawunti yakho.Le miyalezo ingabuyekezwa kamuva futhi ibe yingxenye yomongo osiza ekwenzeni izimpendulo zibe ngezomuntu siqu. Ngokuphambene nalokho, izingxoxo zesikhashana azinawo umjikelezo uma usuvale iseshini: awukwazi ukuziqala kabusha, futhi aziveli ohlwini lwakho lwezingxoxo. Ngokombono womsebenzisi, umqondo ucacile: lokho okushoyo lapho kuhlala lapho.
Usizo olusemthethweni lwe-OpenAI luchaza le modi njengengxoxo eqala "kusukela ekuqaleni".Imodeli ayiboni izingxoxo zakho zangaphambilini noma ifinyelele izinkumbulo zesikhathi esidlule, yize isengasebenzisa imiyalelo yakho eyenzelwe wena uma uyivulile (isibonelo, ithoni yokuphendula noma umongo ojwayelekile ngalokho okwenzayo). Lokhu kuhlukaniswa kukhumbuza imiqondo yokwenza ngokwezifiso efana nokuthi Ubuhlakani Bomuntu Siqu be-Googleezilawula indlela okuthandwayo okusetshenziswa ngayo ngaphandle kokuxuba izimo.
Lokhu kuhlukaniswa kubaluleke kakhulu ngoba i-ChatGPT ifunda ekuxhumaneni kwakhoYonke imibuzo, yonke idokhumenti elayishiwe, kanye nayo yonke imininingwane owabelana ngayo ezingxoxweni ezivamile kungasiza ekulungiseni ukuziphatha kwemodeli ezingxoxweni zakho zesikhathi esizayo. Imizamo yengxoxo yesikhashana "yokuphula" leyo ntambo futhi ivimbele ukushintshana okukodwa ekuthonyeni izimpendulo ezilandelayo noma ukuhlala kubonakala emlandweni wakho wengxoxo.
Kwenzekani ngempela uma usebenzisa ingxoxo yesikhashana?
Ukusebenzisa ingxoxo yesikhashana kudala uchungechunge lwezinguquko ezicacile endleleni iseshini yakho ephathwa ngayo.Nakuba ekuqaleni ukusetshenziswa kufana kakhulu nengxoxo evamile, ngaphansi kukhona umehluko obalulekile ekugcinweni, inkumbulo kanye nokuqeqeshwa kwemodeli.
Uma ingxoxo yesikhashana isebenza, okulandelayo kwenzeka ngokusho kolwazi oluvela ku-OpenAI. kanye nezincazelo esezivele zisakazwe:
- Ingxoxo ayiveli emlandweni wakho wengxoxoNgeke ubone isihloko noma okuqukethwe phakathi kwezingxoxo zakho zakamuva.
- I-ChatGPT "ayikhumbuli" leyo ngxoxo yokuxhumana kwesikhathi esizayoLokho okushoyo lapho akusebenziswa kabusha njengomongo uma ubuyela engxoxweni evamile kamuva.
- Okuqukethwe kwengxoxo yesikhashana akufakiwe kwimemori ye-ChatGPT. (uma unesici sememori esiqhubekayo esivuliwe).
- Lesi sigaba asisetshenziselwa ukuqeqesha amamodeli e-AI unomphela.Ngamanye amazwi, akufanele ifake ohlelweni njengoba kwenza izingxoxo ezijwayelekile uma unikeza imvume.
- I-OpenAI igcina ikhophi yengxoxo kuze kube yizinsuku ezingu-30 ngezizathu zokuphepha nokuthobela imithetho (ukutholwa kokuxhashazwa, ukukhwabanisa, ukusetshenziswa okungekho emthethweni, njll.), futhi ngemva kwaleso sikhathi kuyakususa, indlela efanayo naleyo ye- Izithombe zesikhashana ze-Google.
Leso sikhathi sokukopisha esingafika ezinsukwini ezingu-30 singenye yezinto ezibangela ukukhathazeka okukhulu.Nakuba ungenakukwazi ukuyifinyelela futhi ingaveli ku-interface yakho, isakhona kumaseva enkampani isikhathi esilinganiselwe. Incazelo esemthethweni ivame ukuthi lesi sikhathi siyadingeka ukuze kuphenywe ukusetshenziswa kabi, kuphendulwe ezidingweni zomthetho, futhi kugcinwe ukulandelelwa okuyisisekelo.
Ngesikhathi seseshini, ukuziphatha kufana nokuxoxa "bukhoma" okuvamile.Imodeli ingabhekisela kulokho okushilo imiyalezo embalwa ngaphambilini, iyisebenzise njengomongo osheshayo futhi ikhiqize izimpendulo ezihambisanayo. Kodwa ngokushesha nje lapho uvala noma uphuma engxoxweni yesikhashana, leyo ngxoxo iyanqamuka. Ayikho inketho yokubuyela ezinsukwini kamuva bese uqhubeka lapho ushiye khona.
Iyini ngempela ingxoxo yesikhashana ku-ChatGPT?
Ukusetshenziswa okusobala kakhulu kwengxoxo yesikhashana yilapho udinga ukuxoxa ngezindaba zomuntu siqu, ezibucayi, noma ezibucayi kakhulu. Futhi awukhululekile ukuthi ziqoshwe efayeleni lakho noma zisetshenziswa ukuqeqesha imodeli. Lokhu kufaka phakathi izingxoxo ngempilo, ezezimali, izingxabano zomsebenzi, izinkinga zomndeni, ukuthuthukisa imibono yebhizinisi, njll.
Abasebenzisi abaningi bayisebenzisa ngqo lapho bephoqeleka ukwabelana ngedatha abangathanda ukuthi "ingaqoshwa".Lokhu kungaba idokhumenti equkethe ulwazi oluyimfihlo, umbhalo wezomthetho onamagama afanele, isu lobuchwepheshe eliyimfihlo, noma imininingwane engase ihlonze abantu abathile. Nakuba kungenakwenzeka ukuba kungabikho bungozi, ingxoxo yesikhashana ikuvumela ukuthi unciphise lokho kuvezwa uma kuqhathaniswa nokusebenzisa amakamelo okuxoxa avamile.
Kodwa kunezinye izindlela ezivamile nezisetshenziswa nsuku zonke ezihambisana kahle nale modiIsibonelo, uma uchithe amasonto "uqeqesha" i-ChatGPT ukuthi iphendule ngesitayela esithile sobungcweti, ithoni, noma indlela, ungase ungafuni umbuzo owodwa ohluke ngokuphelele ophazamisa lokho kwenziwa komuntu siqu. Ingxoxo yesikhashana isebenza njengendawo engathathi hlangothi: ubuza umbuzo wakho, uthole impendulo, futhi ayithinti konke okunye okuhlangenwe nakho kwakho.
Kuwusizo futhi uma ufuna izimpendulo ezingathonywanga yizingxoxo zakho zangaphambilini.Uma ubulokhu uxoxa ngesihloko esithile izinsuku eziningana (isibonelo, iphrojekthi yakho yomsebenzi) futhi ngokuzumayo ufuna ukucela usizo ngento ehluke ngokuphelele (indaba yomuntu siqu, umbuzo wezomthetho, inkinga yobuchwepheshe ehlobene kancane), ingxoxo yesikhashana iqinisekisa "ikhanvasi engenalutho" lapho uhlelo lungenalo i-inertia yangaphambili.
Ezindaweni zobungcweti nezebhizinisi, le ndlela yesikhashana inemiphumela eminingi nakakhulu.Amaqembu okuthuthukisa, iminyango yokumaketha, noma abeluleki bangasebenzisa izingxoxo zesikhashana ukwenza izivivinyo ezisheshayo, izivivinyo ze-A/B, ama-ejenti engxoxo esampula, noma basebenze ngokusebenzisa imisebenzi ngaphandle kokuthi lokho okuqukethwe kuxutshwe nedatha "engathi sína" eqeqesha imodeli esetshenziswa ekukhiqizeni.
Lokho ingxoxo yesikhashana engenzi (nokuthi iyini imikhawulo yayo)
Enye yezinto ezivame kakhulu ukungaqondani ukucabanga ukuthi ingxoxo yesikhashana iyindlela yokungaziwa ngokuphelele futhi engenazo izithiyo.Kufana nokuthi ukhuluma nomodeli wendawo okhona kukhompyutha yakho kuphela. Akunjalo. Nakuba ingeza izendlalelo zobumfihlo uma kuqhathaniswa nezingxoxo ezijwayelekile, inemikhawulo ecacile okufanele ikhunjulwe.
Okokuqala, ingxoxo yesikhashana akuyona ingxoxo ongayiqala kabusha kamuva.Kuqondile, cishe "kuyalahleka." Uma ngemva kwesikhashana ufuna ukubuyela kuleso sihloko esifanayo nomongo ogcwele, ayikho indlela yokuqhubeka lapho ugcine khona: kuzodingeka uqale kusukela ekuqaleni kwenye ingxoxo, yesikhashana noma evamile.
Ngeke futhi ikuvumele ukuthi weqe imikhawulo noma imikhawulo yohlelo lwakhoUma i-akhawunti yakho inezilinganiso ezithile zokusetshenziswa, imikhawulo yemilayezo, izifanekiso ezitholakalayo, noma izici, ingxoxo yesikhashana ingaphansi kwaleyo mithetho efanayo. Akuyona indlela yokugwema izilawuli, izihlungi zokuqukethwe, noma izinqubomgomo zokuphepha.
Esinye isici esibalulekile ukuthi, yize ingxoxo yesikhashana ingasetshenziswa ukuqeqesha amamodeliIngxoxo ingase ibuyekezwe ngaphakathi phakathi nalezo zinsuku ezingu-30 (isibonelo, ukuhlolwa kwezokuphepha noma impendulo yesigameko). Lokhu kusho ukuthi akuyona uhlelo oluphephile oluyi-100% lapho kungekho omunye umuntu ongabona khona okubhalile.
Ngaphezu kwalokho, kusalokhu kungaqiniseki ngokuthi kusho ukuthini ngempela “ukungaqeqeshi amamodeli”.Isithembiso esivamile senkampani ukuthi le datha ayifakwanga ekuqeqeshweni okuphelele noma ayisetshenziselwa ukuthuthukisa imodeli yomhlaba wonke. Kodwa-ke, abasebenzisi abaningi bayazibuza ukuthi ingasetshenziswa yini ngokuhlanganiswa, ngendlela engaziwa, noma ekuhlolweni kwangaphakathi. Lokhu kuntuleka kokubonakala kukhulisa ukungathembani.
Icala lomsebenzisi obone idatha yengxoxo yesikhashana iphinda ivele
Enye yezindaba ezidale umsindo omkhulu yilezo zomsebenzisi osebenzise ingxoxo yesikhashana ukuze asebenzele umqondo omusha.Yilokho kanye ayengafuni kuqoshwe noma kuhlotshaniswe nephrofayili yakhe. Ngaleso sikhathi, wacabanga ukuthi lokho akusho kuleso sigaba kuzohlala kuyimfihlo futhi ngeke kulandelwe.
Ezinsukwini kamuva, engxoxweni evamile, waqala ukucela imibono yokuvuselela i-bio yakhe ye-InstagramNgokusho kwe-akhawunti yakhe, uhlelo lwaqala ukuphakamisa imisho nemininingwane ayeyishilo yedwa kuleyo ngxoxo yesikhashana, okwabangela i-alamu. Lapho etshela i-AI ngokusobala ukuthi isebenzisa okuqukethwe okuvela engxoxweni yesikhashana, imodeli yaxolisa, yathi yayifake lolo lwazi enkumbulweni yayo, futhi yathembisa ukususa konke okuhlobene naleyo ngxoxo.
Lokhu kuziphatha kuphakamise umbuzo obalulekile kakhulu: ingabe imodeli yayikhipha ulwazi engxoxweni yesikhashana ngempela? Noma ingabe wayemane “eqamba amanga” futhi enikeza izimpendulo ezijwayelekile umsebenzisi azihumusha njengezinga lezinto ezenzeka ngokuzenzakalelayo? Iqembu elithintekile lavuma ukuthi incazelo esemthethweni yayikhuluma ngephutha lemodeli (ukuqamba amanga) hhayi ukwephulwa ngamabomu komsebenzi wesikhashana.
Ekhathazekile, waxhumana nosekelo lwe-OpenAI.Impendulo ayitholile yayiwukuthi, ngokwezinhlelo zabo, ukususa izingxoxo nokusebenzisa imodi yesikhashana kusho ukuthi idatha ayilondolozwa futhi ayisetshenziswa ukuqeqesha amamodeli. Ithimba lokusekela liphakamise ukudideka okungenzeka futhi lachaza ukuthi imodeli ingase ibe nolwazi oluxubile noma imininingwane eqanjiwe, into evamile kuma-LLM lapho ezama ukuba wusizo ngaphandle kwedatha eyanele.
Nokho, umsebenzisi wayengakayiqondi kahle leyo ncazelo.Umbono wakhe wawuwukuthi imininingwane yayicacile kakhulu ukuba ibe yinto eyenzeka ngengozi nje. Abasekeli base bemcela izibonelo eziqondile, izitembu zesikhathi, kanye nezingcezu eziqondile zemiyalezo ukuze bakwazi ukuphenya kabanzi. Futhi lapha amaqiniso amabili aphikisana: isidingo sobufakazi bobuchwepheshe kanye nokwenqaba komsebenzisi ukwabelana ngemininingwane eyimfihlo ayefuna ukuyigcina iyimfihlo.
Impendulo ye-OpenAI kanye nokungabaza okusele
Inqubomgomo esemthethweni ye-OpenAI icacile: idatha yengxoxo yesikhashana ayisetshenziswa ukuqeqesha amamodeli futhi idatha iyasuswa ngemva kwezinsuku ezingu-30.Kodwa-ke, ikhophi ingagcinwa phakathi naleso sikhathi ngenxa yokuphepha, ukuthobela umthetho, kanye nezinjongo zokuthola ukuxhashazwa. Ngokombono womkhiqizo, umyalezo uwukuthi uma ufuna ungqimba olwengeziwe lobumfihlo, sebenzisa isici sengxoxo yesikhashana bese ususa umlando wakho wengxoxo uma ubona kufanele.
Inkinga ukuthi lesi sithembiso singqubuzana nomuzwa wabasebenzisi abaningi wokuthi ukucaca kunqunyelwe.Ayikho indlela elula yokuhlola usuka ngaphandle ukuthi kwenziwani ngaleyo datha. Awukwazi ukungena kwideshibhodi bese ubona ukuthi izingxoxo zakho zesikhashana ziphathwe kanjani, futhi awukwazi ukuqinisekisa ngokuzimela ukuthi azizange zisetshenziswe ngezinjongo zokuqeqesha.
Okwenza izinto zibe zimbi nakakhulu, uhlobo lwamamodeli olimi lwenza kube lula ukuphambanisa ukuhlangana kwezinto ngezinkumbulo.Uma uke washo into ecacile kakhulu kodwa engeyona eyingqayizivele ngokuphelele, imodeli ingase iveze impendulo efanayo kamuva ngenxa yamathuba olimi oluhlanzekile, futhi ubuchopho bakho bungase buchaze ukuthi "buyakhumbula" lokho okushilo, yize empeleni kumane nje kuhlanganisa amaphethini.
Esimweni okukhulunywe ngaso, ukwesekwa kwe-OpenAI kwavuma nokuthi imodeli "yayinemibono" lapho ithi ifake lolo lwazi enkumbulweni yayo.Ngamanye amazwi, i-ChatGPT ngokwayo yaqamba isimangalo sokuthi isebenzisa idatha yengxoxo yesikhashana bese iyisusa. Lolu hlobo lwezimangalo ezingamanga ezivela kumodeli lusebenza njengesikhumbuzo sokuthi izincazelo zayo zendlela esebenza ngayo ngaphakathi azithembekile ngaso sonke isikhathi: ihlala ingumkhiqizi wombhalo, hhayi ithuluzi lokuxilonga elisobala lesistimu yayo.
Ukuqhathaniswa nemodi ye-incognito yesiphequluli
Isifaniso esivame ukuphindwa kakhulu yilesi se-Chrome kanye nemodi yayo edumile ye-incognitookwathi iminyaka eminingi kwathengiswa njengohlobo lwengubo yokungabonakali futhi kamuva kwabonakala kunemikhawulo eminingi kakhulu kunalokho okwakusikiselwa ukumaketha. Lokhu kufana kuholele abaningi ekukhulumeni "ngesikhathi semodi ye-incognito" se-OpenAI.
Kuziphequluli zewebhu, imodi ye-incognito ivimbela idivayisi yakho ekulondolozeni umlando, amakhukhi, noma amafomu.Noma kunjalo, lokhu akuvimbeli umhlinzeki wakho we-inthanethi, iwebhusayithi oyivakashelayo, noma inkampani yakho (uma usenethiwekhi yenkampani) ekuqopheni umsebenzi wakho. Kwenzeka okufanayo ngengxoxo yesikhashana: kuthuthukisa ubumfihlo ngokombono we-akhawunti yakho nomlando, kodwa akusho ukuthi idatha ayidluli kumaseva enkampani noma ukuthi akukho rekhodi lesikhathi.
Umehluko omkhulu ukuthi lapha asigcini nje ngokubeka umlando wethu wokuphequlula engcuphenikodwa futhi nolwazi oluyinkimbinkimbi: imibono yebhizinisi, imibhalo yangaphakathi, amasu ochwepheshe, idatha yomuntu siqu, noma ngisho nezinto eziyimfihlo. Ngakho-ke, okulindelwe ubumfihlo kuphakeme kakhulu, futhi noma yikuphi ukungavumelani kudala ukungathembani okukhulu.
Lokhu kungabaza kuholele abanye abasebenzisi ukuthi bathi ubumfihlo be-OpenAI kanye nokuphepha kwe-inthanethi "kuyayekethisa".Ikakhulukazi uma becela labo abathintekayo ukuthi babelane ngemininingwane ethile ukuze bahlaziye amaphutha angaba khona. Ngokombono womsebenzisi wokugcina, impendulo ethi "sithumelele konke okungahambanga kahle, ngokuningiliziwe" ingangqubuzana neqiniso lokuthi abafuni ngempela ukwabelana ngalolo lwazi nanoma ubani.
Imiphumela yobuchwepheshe neyebhizinisi: iseshini vs isitoreji esiqhubekayo
Ngokombono wezobuchwepheshe, inselele isekuhlukaniseni ngokucacile umongo weseshini nokugcinwa okuqhubekayo.Okungukuthi, ukuhlukanisa phakathi kwalokho imodeli okudingeka ikubone ngesikhathi sangempela ukuze iphendule (lokho oyithumelayo ngaleso sikhathi) nalokho okulondolozwa kamuva njengengxenye yephrofayela noma isethi yedatha yokuqeqesha.
Imikhuba emihle kakhulu kumapulatifomu okuxoxa athuthukile ifaka phakathi ukubethela idatha ezingeni leseshiniSebenzisa izinqubomgomo zokugcina ezilungisekayo futhi ugcine ukulandelelwa kwemicimbi. Lokhu kuvumela okuhlangenwe nakho okwenziwe ngokwezifiso kakhulu ngesikhathi sengxoxo ethile ngaphandle kokuthi ulwazi olubucayi luhlanganiswe nedatha ehlala njalo yomsebenzisi noma yesistimu.
Ebhizinisini, izingxoxo zesikhashana ziwusizo kakhulu kuma-ejenti okusekela. Kudingeka bakhumbule okwenzekayo ekuxhumaneni okuthile (isibonelo, isigameko ikhasimende elisichazayo) ngaphandle kokuthi idatha ifakwe ngokungalawuleki "engqondweni" yenkampani iyonke. Lokhu kufaka phakathi nobufakazi bomqondo, izinkomba zangaphakathi, kanye nezibonelo ezihloselwe ukuhlukaniswa nendawo yokukhiqiza.
Izinkampani ezakha izinhlelo zokusebenza ezenziwe ngokwezifiso zivame ukuhlanganisa izinsizakalo zamafu nesitoreji sokuqukethwe sesikhashanaAmapayipi okufihla ulwazi kanye nezinhlelo zokuhlola nawo abalulekile. Umgomo uwukuba i-AI inikeze izimpendulo ezihambisanayo kulo lonke iseshini ngaphandle kokudala ingobo yomlando engapheli yakho konke okushiwoyo. Ukuhlanganisa lokhu namadeshibhodi, amathuluzi okuhlaziya, ngisho ne-BI kudinga ukuklama ukwakheka ngokucabangela kokubili ukukhuliswa kanye nokuphathwa kwedatha.
Okwamanje, ezokuphepha kanye nemithethonqubo kubeka izinhloso zaboUkuhlolwa kokuphepha kwe-inthanethi, ukuhlolwa kokungena, ukuqina kohlelo, kanye nokulawula ukufinyelela okuqhubekayo kubalulekile lapho kuphathwa idatha ebucayi, noma ngabe ngokwembono "iyinto yesikhashana." Ukwehluleka kwesendlalelo sokuphepha kungashintsha ingxoxo yesikhashana ibe umthombo wokuvuza okukhulu njenganoma iyiphi i-database ehlala njalo.
Ingabe konke kuyanyamalala ngempela ngemva kwezinsuku ezingu-30?
Omunye wemibuzo emikhulu abantu abaningi abayibuzayo ukuthi ngabe kukhona yini ubufakazi obuqinile bokuthi idatha yengxoxo yesikhashana isusiwe ngokuphelele. Ngemva kwalezo zinsuku ezingu-30, azisetshenziswa “nganoma iyiphi indlela” ngemva kwalokho. Okwamanje, impendulo eqotho iwukuthi sithembele ekuthembekeni ezinqubweni zangaphakathi zenkampani kanye nemithetho edinga ukuba ihambisane nazo.
Ayikho indlela yomphakathi yokuhlola ngqo lokho kususwa ku-akhawunti yomsebenzisiAwukwazi ukuvula ideshibhodi ekhombisa ukuthi ingxoxo ngayinye ihlanzwe kanjani, futhi akukho umbiko ongalandwa ochaza ukubhujiswa kwedatha. Empeleni, lesi isithembiso sesivumelwano: I-OpenAI ithi iyakwenza, futhi uma ingakwenza, izobhekana nezinkinga ezinkulu zomthetho kanye nedumela.
Kubasebenzisi abakhathazeke kakhulu ngobumfihlo, lokhu kungase kunganele.Eqinisweni, abanye abantu bangathanda ukusebenzisa amamodeli endawo kuphela (ama-LLM asebenza kukhompyutha yabo) ukuze babe nokulawula okuphelele lapho idatha yabo igcinwa khona. Inkinga, njengoba abaningi besho, ukuthi ihadiwe yasekhaya ayihlali ikwazi ukwenza umsebenzi: amamodeli anamandla kakhulu adinga imishini enzima kakhulu, futhi emishinini ephansi, ulwazi luba luncane futhi lunciphe.
Okwamanje, i-ChatGPT isalokhu iyinketho enikeza izimpendulo zekhwalithi engcono kakhulu kwabaningi.Lokhu kudinga ibhalansi ethile: ukuyisebenzisa kulokho okwengeza inani elikhulu, kodwa ngokubhukha mayelana nokwabelana ngolwazi abangafuni ukulushiya endaweni yabo eseduze. Ingxoxo yesikhashana inciphisa okunye kwalokho kwesaba, kodwa ayikususi ngokuphelele.
Njengoba kunikezwe leli qiniso, isincomo esisebenzayo ngokuvamile ukusebenzisa ingxoxo yesikhashana ukunciphisa ukuvezwa.Gwema ukuthumela idatha ongeke uyabelane ngayo nanoma iyiphi enye isevisi eku-inthanethi, futhi uma izinga lokuzwela liphezulu kakhulu, cabanga ngokungathi sína ngezixazululo ezisendaweni noma amamodeli endawo, noma ngabe ahilela izindleko eziphakeme kanye nokusebenza okubi kakhulu.
Ngamafuphi, ingxoxo yesikhashana ye-ChatGPT iyithuluzi eliwusizo lokuhlukanisa izingxoxo ezibucayi ekusebenziseni kwakho kwansuku zonke.Kodwa-ke, akusona isiqinisekiso esiphelele sokungaziwa noma isibhamu somlingo sazo zonke izinkinga zobumfihlo. Kuyasiza ukuvimbela izihloko ezithile ukuthi zingathonyi izingxoxo zakho zesikhathi esizayo, zigcinwe emlandweni wakho, futhi ziqeqeshe imodeli, ngezindleko zokwamukela ukuthi zizohlala kumaseva e-OpenAI kuze kube yizinsuku ezingu-30 ngaphansi kwemithetho yazo yokuphepha nokuthobela imithetho.
