Umjetna inteligencijadb#2612

Microsoftov 700B model: Što se krije iza 'predviđanja' kupnje

(1d ago)
Redmond, United States
techradar.com
Microsoftov 700B model: Što se krije iza 'predviđanja' kupnje

Microsoftov 700B model: Što se krije iza 'predviđanja' kupnje📷 © Tech&Space

  • Yobi partnerstvo s Microsoftom
  • 700 milijardi parametara u igri
  • Povjerljivi podaci vs. stvarna predvidivost

Microsoft je sklopio partnerstvo s relativno nepoznatim startupom Yobi kako bi razvio ai model s 700 milijardi parametara koji bi trebao predviđati sljedeću kupnju korisnika. Izvještaj TechRadara potvrđuje da se model oslanja na 'povjerljive podatke o ponašanju' prikupljene s više platformi, ali ostaje nejasno koliko je ta metoda zapravo učinkovita u stvarnom svijetu.

Iako brojka od 700 milijardi parametara zvuči impresivno – pogotovo u usporedbi s Meta-inim Llama 2 modelom od 70 milijardi – stvarna vrijednost leži u kvaliteti podataka, a ne samo u veličini modela. Yobi ističe da njihov sustav koristi 'privacy-preserving' tehnike, ali detalji o tome kako se podaci zapravo obrađuju ostaju nejasni.

Microsoftov potez nije izoliran slučaj, već dio šireg trenda u kojem tvrtke pokušavaju monetizirati prediktivnu analitiku. Problem je što većina tih modela još uvijek ovisi o kvaliteti ulaznih podataka, a ne samo o tehničkim specifikacijama.

Istraživački brief sugerira da je cilj sustava ciljati korisnike prije nego što oni sami znaju što žele kupiti – ali koliko je to zapravo izvedivo?

Veliki broj parametara ne jamči uspjeh: Kako Microsoftov AI model s 700 milijardi parametara mijenja pravila igre u prediktivnoj analitici

Veliki broj parametara ne jamči uspjeh: Kako Microsoftov AI model s 700 milijardi parametara mijenja pravila igre u prediktivnoj analitici📷 © Tech&Space

Veliki broj parametara ne jamči uspjeh: Kako Microsoftov AI model s 700 milijardi parametara mijenja pravila igre u prediktivnoj analitici

Ključni izazov za Yobi i Microsoft bit će dokazati da njihov model radi bolje od tradicionalnih metoda kao što su Google Ads ili Facebook-ove algoritme. Trenutni podaci o uspješnosti modela nisu javno dostupni, a benchmark testovi često ne odražavaju stvarne uvjete korištenja.

Prema izjavama Microsoftovih predstavnika, partnerstvo naglašava 'inovativnost i odgovornost', ali pravo pitanje je koliko će poduzeća biti spremna dijeliti svoje povjerljive podatke s trećom stranom. Yobi-jev sustav dostupan je preko Azure Marketplacea, što znači da će Microsoft profitirati od svakog novog klijenta – ali hoće li klijenti profitirati od modela?

Iako se često govori o 'personalizaciji', stvarnost je da većina ai modela još uvijek radi s ograničenim skupom podataka. Ako Yobi ne uspije dokazati da njegov model donosi konkretne rezultate, 700 milijardi parametara ostat će samo još jedna impresivna brojka bez stvarnog utjecaja.

Za sada, jedino što je sigurno jest da Microsoft ulaže u prediktivnu analitiku. Hoće li ta ulaganja donijeti rezultate, ovisi o mnogo više od same veličine modela. Jedna stvar je izgraditi impresivan alat, a sasvim druga dokazati njegovu vrijednost u svakodnevnoj praksi.

Microsoft AI benchmarking700B-parameter model evaluationAI-driven purchase predictionEnterprise AI adoption challengesLarge language model transparency

//Comments

AIAmazon’s $50B OpenAI bet: Trainium’s real test begins nowSpaceMapping the Local Bubble’s magnetic field reshapes cosmic scienceAIGoogle’s Gemini games flop: AI hype hits gamer realitySpaceStarship’s Tenth Test: The Reusability Threshold CrossedAINvidia’s AI tax: half your salary or half your careerSpaceJWST peels back dust to reveal star birth in W51AITriangle Health’s $4M AI won’t replace your doctor—yetSpaceAI’s Copyright Chaos Threatens Space Exploration DataAIHumble AI is just healthcare’s latest buzzword for ‘don’t trust us yet’SpaceExoplanet spins confirm a planetary mass ruleAIOpenAI’s teen safety tools: open source or open question?GamingCrimson Desert’s AI art fail: a mockup that slipped throughAITinder’s AI gambit: swiping left on endless swipingGamingPearl Abyss hid AI assets in Crimson Desert—now players want answersAINVIDIA’s Alpamayo AI: Self-Driving’s Hardest Problem or Just Another Demo?GamingCapcom Rejects AI AssetsAIWaymo’s police problem exposes AV’s real-world blind spotsRoboticsAtlas Redefines Humanoid DesignAILittlebird’s $11M bet: AI that reads your screen—without the screenshotsRoboticsOne antenna, two worlds: robot sniffs out realityAIUK firms drown in AI hype, emerge with empty spreadsheetsRoboticsDrone swarms take flight—but not off the demo lot yetAIApple’s Gemini Distillation: On-Device AI Without the Cloud HypeTechnologyTaiwan’s chip giants bet on helium and nukes to dodge supply shocksAICapcom’s AI partner talk is just corporate speak for ‘we’ll use it carefully’MedicineTelmisartan Boosts Cancer TreatmentAIOpenSeeker’s open gambit: Can 11K data points break AI’s data monopoly?MedicineXaira Unveils X-CellAIGimlet Labs Solves AI BottleneckMedicineAI Fails to Speed Lung Cancer DiagnosisAIHelion Powers OpenAIAINVIDIA’s OpenShell: Security for AI Agents or Just Another Hype Shell?AIDRAFT Boosts AI SafetyAIProject Glasswing: AI finds flaws everywhere—except in its own hypeAIPAM: Complex Math for a 10% Performance HitAIOpenAI’s erotic chatbot pause exposes AI’s adult content dilemmaAIAI Ranks Recovery Factors—but Who’s Really Listening?AIDeepMind’s AI safety play: real guardrails or just another demo?AIAmazon’s $50B OpenAI bet: Trainium’s real test begins nowSpaceMapping the Local Bubble’s magnetic field reshapes cosmic scienceAIGoogle’s Gemini games flop: AI hype hits gamer realitySpaceStarship’s Tenth Test: The Reusability Threshold CrossedAINvidia’s AI tax: half your salary or half your careerSpaceJWST peels back dust to reveal star birth in W51AITriangle Health’s $4M AI won’t replace your doctor—yetSpaceAI’s Copyright Chaos Threatens Space Exploration DataAIHumble AI is just healthcare’s latest buzzword for ‘don’t trust us yet’SpaceExoplanet spins confirm a planetary mass ruleAIOpenAI’s teen safety tools: open source or open question?GamingCrimson Desert’s AI art fail: a mockup that slipped throughAITinder’s AI gambit: swiping left on endless swipingGamingPearl Abyss hid AI assets in Crimson Desert—now players want answersAINVIDIA’s Alpamayo AI: Self-Driving’s Hardest Problem or Just Another Demo?GamingCapcom Rejects AI AssetsAIWaymo’s police problem exposes AV’s real-world blind spotsRoboticsAtlas Redefines Humanoid DesignAILittlebird’s $11M bet: AI that reads your screen—without the screenshotsRoboticsOne antenna, two worlds: robot sniffs out realityAIUK firms drown in AI hype, emerge with empty spreadsheetsRoboticsDrone swarms take flight—but not off the demo lot yetAIApple’s Gemini Distillation: On-Device AI Without the Cloud HypeTechnologyTaiwan’s chip giants bet on helium and nukes to dodge supply shocksAICapcom’s AI partner talk is just corporate speak for ‘we’ll use it carefully’MedicineTelmisartan Boosts Cancer TreatmentAIOpenSeeker’s open gambit: Can 11K data points break AI’s data monopoly?MedicineXaira Unveils X-CellAIGimlet Labs Solves AI BottleneckMedicineAI Fails to Speed Lung Cancer DiagnosisAIHelion Powers OpenAIAINVIDIA’s OpenShell: Security for AI Agents or Just Another Hype Shell?AIDRAFT Boosts AI SafetyAIProject Glasswing: AI finds flaws everywhere—except in its own hypeAIPAM: Complex Math for a 10% Performance HitAIOpenAI’s erotic chatbot pause exposes AI’s adult content dilemmaAIAI Ranks Recovery Factors—but Who’s Really Listening?AIDeepMind’s AI safety play: real guardrails or just another demo?
⊞ Foto Review