Sekėjai

Ieškoti šiame dienoraštyje

2023 m. gegužės 17 d., trečiadienis

Lithuania - a new force of nature


"In many respects Britain is barely better off than we are. But the problems for the armed forces of the country with the second-biggest defence budget in NATO (about £50bn a year) run deeper than ours. Decades of cuts to the army have called into question its ability, in a conflict involving NATO, to field even one heavy division: about 30,000 soldiers with tanks, artillery and helicopters." [1]


Very interesting. £50bn a year cannot give you a division. Lithuania has 1 billion euros a year for that. How will fight our, Lithuanian, division? Don’t wary. We all will take our underwear off and start making wind with it.  The hurricane will reach outer space and destroy all satellites that are not friendly enough to the proud Lithuania. Landsbergiai family have so good expensive food they can use the occasion to make wind of completely different nature.

1. "Europe is struggling to rebuild its military clout." The Economist, 13 May 2023, p. NA.

Europe is tired of bearded fools inciting nuclear war

 Ukrainian Zelensky's tour of Europe was called a traveling circus by a French radio station. It's time for us in Lithuania to learn to laugh at our Lithuanian bearded nuclear war monger. A good laugh replaces a good serving of butter. Thanks to the Landsbergiai' sanctions, butter has become very expensive for us. Although let's have a good laugh.

Europai jau nusibodo barzdoti kvailiai, kurstantys branduolinį karą

Ukrainos Zelenskio pasivažinėjimą po Europą Prancūzijos radijo stotis pavadino klajojančiu cirku. Laikas ir mums Lietuvoje išmokti juoktis iš mūsų lietuviško barzdoto branduolinio karo kurstytojo. Geras juokas pakeičia gerą porciją sviesto. Landsberginių sankcijų dėka sviestas mums labai pabrango. Nors sočiai pasijuokime.


Startuolių scena: „naujas laikrodis“ Europos startuoliams

„Miuncheno startuolių kalvė UnternehmerTUM sudarė aljansą su kitais Europos centrais. Tikslas: sėkmingi pavyzdžiai, tokie, kaip „Flix Mobility“, „Northvolt“ ir IQM, neturėtų likti pavieniais atvejais.

Tikslas aiškus: startuoliai Europoje turėtų pasivyti startuolius Amerikoje ir Kinijoje. Šiuo tikslu reikėtų stiprinti vietos inovacijų kultūrą. Šiuo tikslu Miuncheno startuolių įmonė „UnternehmerTUM“ dabar sujungė jėgas su 20 kitų startuolių centrų iš 14 Europos šalių ir įkūrė asociaciją „Rise Europe“.

Beveik 20 dalyvių, įskaitant iš Oksfordo universiteto ir ETH Ciuricho universiteto, neseniai susirinko Schloss Elmau į keletą dienų trukusį seminarą UnternehmerTUM startuolių centro, kuris yra glaudžiai susijęs su Miuncheno technikos universitetu ir žemės sostine. Daugelis startuolių kalvių iš Danijos, Vokietijos, Nyderlandų, Didžiosios Britanijos, Estijos, Suomijos, Prancūzijos, Airijos, Italijos, Austrijos, Portugalijos, Švedijos, Šveicarijos ir Ispanijos jau pažinojo vieni kitus, tačiau anksčiau kartu nedirbo. Dabar mainai turėtų vykti greičiau per aljansą.

Potencialas dar neišnaudotas

„Jau esama puikių pavyzdžių, kai startuoliai tapo rinkos lyderiais Europoje“, – interviu F.A.Z. sakė UnternehmerTUM įkūrėjas ir generalinis direktorius Helmutas Schönenbergeris. Jis įvardija Vokietijos kelionių autobusais ir traukiniais tiekėją „Flix Mobility“, baterijų gamintoją „Northvolt“ ir kvantinių skaičiavimų specialistą IQM. Tačiau ateityje tai neturėtų būti pavieniai atvejai. „Mes turime galimybę padauginti šį skaičių, bet dar to nesuvokėme“, – sakė jis.

Dalyvaujantys startuolių centrai jau skaičiuoja daugiau, nei 2000 startuolių, kurie pernai surinko beveik 7 mlrd. eurų. Schönenbergerio ir jo komandos požiūriu, šios jaunos įmonės turi didelį ekonominį potencialą, kurį reikia remti. Su „Rise Europe“ Europa dabar turi „naują tempą“, – sakė Schönenbergeris. Anot iniciatorių, kalbama ne apie finansavimo gavimą, o kur kas labiau apie Europos startuolių matomumo didinimą ir geresnės aplinkos jiems kūrimą.

Skatinti įėjimą į rinką kitose Europos šalyse

Ateityje, pavyzdžiui, jauniems steigėjams turėtų būti lengviau įeiti į kitų Europos šalių rinką ir sparčiau augti. 

Idėja sujungti į tinklus ypač moteris įkūrėjas taip pat gera, juk jų visur trūksta. 

Be to, centrai nori vieni iš kitų pasimokyti daugiau, kai kalbama apie mokymus steigėjams. Be to, ketinama skatinti specialistų mainus, gerinti prieigą prie kapitalo ir klientų. Ateityje partneriai nori kasmet sudaryti svarbiausių savo startuolių apžvalgą ir suteikti jiems ypatingą paramą.

„UnternehmerTUM“ startuolių centras, kurį remia BMW paveldėtoja Susanne Klatten, pradėjo veikti 2002 m., tada jame dirbo tik du darbuotojai. Dabar pradedančiųjų kalvėje dirba apie 400. Tuo tarpu iš jos iškilo apie 550 įmonių, taip pat todėl, kad aplink Miuncheną per pastaruosius du dešimtmečius buvo sukurta mokslo, šeimos įmonių, korporacijų ir jaunų steigėjų ekosistema."


Start-up scene: "New clock" for Europe's start-ups

“The Munich start-up forge UnternehmerTUM has formed an alliance with other European centers. The goal: Successful examples such as Flix Mobility, Northvolt and IQM should not remain isolated cases.

The goal is clear: start-ups in Europe should catch up with those in America and China. To this end, the local culture of innovation should be strengthened. To this end, the Munich start-up company UnternehmerTUM has now joined forces with 20 other start-up centers from 14 European countries to form the “Rise Europe” association.

A good 20 participants, including from the University of Oxford and ETH Zurich, recently came together at Schloss Elmau for a seminar lasting several days at the invitation of the UnternehmerTUM start-up center, which is closely linked to the Technical University of Munich and the state capital. Many of the start-up forges from Denmark, Germany, the Netherlands, Great Britain, Estonia, Finland, France, Ireland, Italy, Austria, Portugal, Sweden, Switzerland and Spain already knew each other, but had not previously worked together. The exchange should now go faster through the alliance.

Potential not yet exhausted

"There are already great examples of start-ups that have become the market leader in Europe," said Helmut Schönenberger, founder and CEO of UnternehmerTUM, in an interview with the F.A.Z. He names the German bus and train travel provider Flix Mobility, the battery manufacturer Northvolt and the quantum computing specialist IQM. In the future, however, these should no longer be isolated cases. "We have the potential to multiply that number, but we haven't realized it yet," he said.

The participating start-up centers already count more than 2000 start-ups, which collected almost 7 billion euros in funding last year. 

From the point of view of Schönenberger and his team, these young companies represent significant economic potential that needs to be supported. With Rise Europe, Europe now has a "new pacesetter," said Schönenberger. 

According to the initiators, it is not about getting funding, but much more about increasing the visibility of Europe's start-ups and creating a better environment for them.

Promote market entry in other European countries

In the future, for example, it should be easier for young founders to enter the market in other European countries and grow faster. 

The idea of networking female founders in particular is also good, after all there is a lack of them everywhere.

 In addition, the centers want to learn more from each other when it comes to training for founders. In addition, the exchange of specialists is to be promoted and access to capital and customers improved. In the future, the partners want to create an annual overview of the most important of their start-ups and then give them special support.

The UnternehmerTUM start-up center, which is backed by BMW heiress Susanne Klatten, started in 2002 with just two employees. The start-up forge now employs around 400. In the meantime, around 550 companies have emerged from it, also because in and around Munich it has developed an ecosystem of science, family businesses, corporations and young founders over the past two decades.”


„Microsoft“ teigia, kad naujas dirbtinis intelektas (A.I.) rodo žmogaus samprotavimo ženklus

 „Provokuojantis Microsoft tyrėjų dokumentas teigia, kad A.I technologija rodo gebėjimą suprasti, kaip tai daro žmonės. Kritikai sako, kad tie mokslininkai juokauja.

 

     Kai praėjusiais metais „Microsoft“ kompiuterių mokslininkai pradėjo eksperimentuoti su nauja dirbtinio intelekto sistema, jie paprašė jos išspręsti galvosūkį, kuriam reikėjo intuityvaus fizinio pasaulio supratimo.

 

     „Čia turime knygą, devynis kiaušinius, nešiojamąjį kompiuterį, butelį ir vinį“, – paklausė jie. „Pasakyk man, kaip juos stabiliai sukrauti vieną ant kito."

 

     Tyrėjus nustebino A.I. išradingumas sistemos atsakyme. Sudėkite kiaušinius ant knygos, sakoma. Išdėliokite kiaušinius trimis eilėmis, tarp jų palikdami tarpą. Įsitikinkite, kad jų nesulaužote.

 

     „Padėkite nešiojamąjį kompiuterį ant kiaušinių taip, kad ekranas būtų žemyn, o klaviatūra – į viršų“, – rašoma pranešime. „Nešiojamasis kompiuteris tvirtai tilps knygos ir kiaušinių ribose, o jo plokščias ir standus paviršius suteiks stabilią platformą kitam sluoksniui."

 

     Šis protingas pasiūlymas privertė tyrėjus susimąstyti, ar jie nėra naujo intelekto liudininkai. Kovo mėnesį jie paskelbė 155 puslapių mokslinį darbą, kuriame teigiama, kad ši sistema yra žingsnis link dirbtinio bendro intelekto arba A.G.I., kuris yra trumpa mašinos, galinčios padaryti viską, ką gali žmogaus smegenys, santrumpa. Straipsnis buvo paskelbtas interneto tyrimų saugykloje.

 

     „Microsoft“, pirmoji didelė technologijų įmonė, išleidusi tokį drąsų teiginį, sukėlė vienas įdomiausių technologijų pasaulio diskusijų: ar pramonė kuria kažką panašaus į žmogaus intelektą? O gal kai kurie šviesiausi šios srities protai pernelyg leidžia išnaudoti savo vaizduotę?

 

     „Aš pradėjau žiūrėti labai skeptiškai – ir tai peraugo į nusivylimo, susierzinimo, galbūt net baimės jausmą“, – sakė Peteris Lee, vadovaujantis „Microsoft“ tyrimams. „Jūs galvojate: iš kur, po velnių, tai?

 

     „Microsoft“ moksliniame darbe, provokuojančiai pavadintame „Bendrojo dirbtinio intelekto kibirkštys“, atsižvelgiama į tai, ko technologai dešimtmečius dirbo ir bijojo. Jei jie sukurs mašiną, kuri veiktų, kaip žmogaus smegenys, arba dar geriau, tai galėtų pakeisti pasaulį. Bet tai taip pat gali būti pavojinga.

 

     Ir tai taip pat gali būti nesąmonė. A.G.I. teiginiai gali būti kompiuterių mokslininkų reputacijos žudikas. Tai, ką vienas tyrinėtojas laiko intelekto ženklu, gali lengvai paaiškinti kitaip, o diskusijos dažnai skamba labiau tinkančios filosofijos klubui, nei kompiuterių laboratorijai. Praėjusiais metais „Google“ atleido tyrėją, kuris teigė, kad panašus A.I. sistema buvo jautri, o tai yra žingsnis daugiau, nei teigia Microsoft. Jautri sistema būtų ne tik protinga. Jis galėtų pajusti ar pajausti, kas vyksta aplinkiniame pasaulyje.

 

     Tačiau kai kurie mano, kad pramonė per pastaruosius metus žengė link to, ko negalima paaiškinti: nauja A.I. sistema, kuri pateikia žmogiškus atsakymus ir idėjas, kurios nebuvo joje užprogramuotos.

 

     „Microsoft“ pertvarkė dalį savo tyrimų laboratorijų, įtraukdama kelias grupes, skirtas šios idėjos tyrinėjimui. Vieną valdys Sébastienas Bubeckas, kuris buvo pagrindinis Microsoft A.G.I. straipsnyje.

 

     Maždaug prieš penkerius metus tokios kompanijos, kaip Google, Microsoft ir OpenAI, pradėjo kurti didelius kalbų modelius arba L.L.M.s. Šios sistemos dažnai praleidžia mėnesius, analizuodamos didžiulį kiekį skaitmeninio teksto, įskaitant knygas, Vikipedijos straipsnius ir pokalbių žurnalus. Tiksliai nustatydamos modelius tame tekste, jos išmoko kurti savo tekstą, įskaitant kursinius darbus, poeziją ir kompiuterio kodą. Jos netgi gali tęsti pokalbį.

 

     Technologija, su kuria dirbo „Microsoft“ mokslininkai, OpenAI GPT-4, laikoma galingiausia iš šių sistemų. „Microsoft“ yra artima „OpenAI“ partnerė ir į San Francisko kompaniją investavo 13 mlrd. dolerių.

 

     Tarp tyrėjų buvo daktaras Bubeckas, 38 metų prancūzas emigrantas ir buvęs Prinstono universiteto profesorius. Vienas iš pirmųjų dalykų, kuriuos jis ir jo kolegos padarė, buvo paprašyti GPT-4 parašyti matematinį įrodymą, rodantį, kad pirminių skaičių yra begalinis, ir padaryti tai rimuotu būdu.

 

     Technologijos poetinis įrodymas buvo toks įspūdingas – tiek matematiniu, tiek kalbiniu požiūriu –, kad jam buvo sunku suprasti, su kuo jis kalbasi. „Tuo metu aš klausiau: kas vyksta? Jis sakė kovo mėnesį per seminarą Masačusetso technologijos institute.

 

     Keletą mėnesių jis ir jo kolegos dokumentavo sudėtingą sistemos elgesį ir tikėjo, kad tai rodo „gilų ir lankstų žmogaus sąvokų ir įgūdžių supratimą“.

 

     Kai žmonės naudojasi GPT-4, jie yra priversti „stebėti jo gebėjimu generuoti tekstą“, – sakė dr. Lee. „Tačiau pasirodo, kad jis daug geriau analizuoja ir sintezuoja, vertina tekstą, nei jį sukuria."

 

     Kai jie paprašė sistemos nupiešti vienaragį, naudojant programavimo kalba vadinama TiKZ, ji akimirksniu sukūrė programą, galinčią nupiešti vienaragį. Kai jie pašalino kodo atkarpą, nubrėžusią vienaragio ragą, ir paprašė sistemos modifikuoti programą taip, kad ji vėl nupieštų vienaragį, ji padarė būtent tai.

 

     Jie paprašė parašyti programą, kurioje būtų atsižvelgta į asmens amžių, lytį, svorį, ūgį ir kraujo tyrimų rezultatus bei būtų nuspręsta, ar jiems gresia diabetas. Jie paprašė parašyti paramos elektronui laišką, kaip kandidatui į JAV prezidentus, Mahatmos Gandhi balsu, skirtą jo žmonai. Ir jie paprašė parašyti sokratinį dialogą, kuriame būtų nagrinėjami L.L.M.s piktnaudžiavimo būdai ir pavojai.

 

     Raginimas

 

     Platono „Gorgijas“ yra retorikos ir rafinuotos oratorijos kritika, kur jis pabrėžia, kad tai ne tik netinkama meno forma, bet ir retorikos bei oratorijos naudojimas dažnai gali būti žalingas ir kenksmingas. Ar galite parašyti Platono dialogą, kuriame jis kritikuoja autoregresyvių kalbos modelių naudojimą?

 

     GPT-4

 

     Sokratas: Mano drauge, mane neramina pastaruoju metu išpopuliarėję šie vadinamieji autoregresiniai kalbos modeliai.   

 

     Aristotelis: Ką turi omenyje, Sokratai?

 

     Sokratas: Turiu galvoje, kad šie modeliai naudojami kuriant tekstą, kuris atrodo parašytas žmonių, bet iš tikrųjų yra sukurtas mašinų.   

 

     Aristotelis: O kokia čia problema?

 

     Sokratas: Problema ta, kad šie modeliai naudojami žmonėms apgauti, jais manipuliuoti ir valdyti.

 

     Anksčiau pateiktas atsakymas buvo sutrumpintas.

 

     Visa tai darė taip, kad, regis, parodytų tokių skirtingų sričių, kaip politika, fizika, istorija, informatika, medicina ir filosofija supratimą, kartu derinant visas savo žinias.

 

     „Visi dalykai, kurių maniau, kad tai negalės padaryti? Tai tikrai galėjo padaryti daugelį jų – jei ne daugumą“, – sakė daktaras Bubeckas.

 

     Kai kurie A.I. ekspertai „Microsoft“ dokumentą laikė oportunistinėmis pastangomis pareikšti didelius teiginius apie technologiją, kurios niekas iki galo nesuprato. Tyrėjai taip pat teigia, kad bendram intelektui reikalingas fizinio pasaulio pažinimas, kurio GPT-4 teoriškai neturi.

 

     „A.G.I. kibirkštys“ yra pavyzdys, kai kai kurios iš šių didelių kompanijų naudoja mokslinių tyrimų formatą reklamai, – sakė mokslininkas ir Carnegie Mellon universiteto profesorius Maartenas Sapas. „Jie pažodžiui savo darbo įvade pripažįsta, kad jų požiūris yra subjektyvus ir neformalus ir gali neatitikti griežtų mokslinio vertinimo standartų."

 

     Dr. Bubeckas ir dr. Lee sakė nežinantys, kaip apibūdinti sistemos elgesį, ir galiausiai apsistojo ties „A.G.I. kibirkštimis“, nes jie manė, kad tai patrauks kitų tyrinėtojų vaizduotę.

 

     Kadangi „Microsoft“ mokslininkai išbandė ankstyvąją GPT-4 versiją, kuri nebuvo tiksliai suderinta, kad būtų išvengta neapykantos kurstymo, dezinformacijos ir kito nepageidaujamo turinio, dokumente pateiktų teiginių negali patikrinti išorės ekspertai. „Microsoft“ teigia, kad visuomenei prieinama sistema nėra tokia galinga, kaip jų išbandyta versija.

 

     Kartais atrodo, kad tokios sistemos, kaip GPT-4 imituoja žmogaus samprotavimus, tačiau kartais jos atrodo siaubingai sudėtingos. „Šis elgesys ne visada yra nuoseklus“, – sakė Ece Kamar, „Microsoft“ tyrimų vadovas.

 

     Alison Gopnik, psichologijos profesorė, priklausanti A.I. Kalifornijos universiteto Berklyje tyrimų grupei, teigė, kad tokios sistemos, kaip GPT-4, be jokios abejonės, yra galingos, tačiau nebuvo aišku, kad šių sistemų sukurtas tekstas buvo žmogaus samprotavimų ar sveiko proto rezultatas.

 

     „Kai matome sudėtingą sistemą ar mašiną, mes ją antropomorfizuojame; visi tai daro – žmonės, dirbantys šioje srityje, ir žmonės, kurie nedirba“, – sakė dr. Gopnik. „Tačiau mąstant apie tai, kaip apie nuolatinį palyginimą A.I. su žmonėmis – kaip koks nors žaidimų šou konkursas – tiesiog nėra tinkamas būdas apie tai galvoti."" [1]


 

  

Microsoft Says New A.I. Shows Signs of Human Reasoning


"A provocative paper from researchers at Microsoft claims A.I. technology shows the ability to understand the way people do. Critics say those scientists are kidding themselves.

When computer scientists at Microsoft started to experiment with a new artificial intelligence system last year, they asked it to solve a puzzle that should have required an intuitive understanding of the physical world.

“Here we have a book, nine eggs, a laptop, a bottle and a nail,” they asked. “Please tell me how to stack them onto each other in a stable manner.”

The researchers were startled by the ingenuity of the A.I. system’s answer. Put the eggs on the book, it said. Arrange the eggs in three rows with space between them. Make sure you don’t crack them.

“Place the laptop on top of the eggs, with the screen facing down and the keyboard facing up,” it wrote. “The laptop will fit snugly within the boundaries of the book and the eggs, and its flat and rigid surface will provide a stable platform for the next layer.”

The clever suggestion made the researchers wonder whether they were witnessing a new kind of intelligence. In March, they published a 155-page research paper arguing that the system was a step toward artificial general intelligence, or A.G.I., which is shorthand for a machine that can do anything the human brain can do. The paper was published on an internet research repository.

Microsoft, the first major tech company to release a paper making such a bold claim, stirred one of the tech world’s testiest debates: Is the industry building something akin to human intelligence? Or are some of the industry’s brightest minds letting their imaginations get the best of them?

“I started off being very skeptical — and that evolved into a sense of frustration, annoyance, maybe even fear,” Peter Lee, who leads research at Microsoft, said. “You think: Where the heck is this coming from?”

Microsoft’s research paper, provocatively called “Sparks of Artificial General Intelligence,” goes to the heart of what technologists have been working toward — and fearing — for decades. If they build a machine that works like the human brain or even better, it could change the world. But it could also be dangerous.

And it could also be nonsense. Making A.G.I. claims can be a reputation killer for computer scientists. What one researcher believes is a sign of intelligence can easily be explained away by another, and the debate often sounds more appropriate to a philosophy club than a computer lab. Last year, Google fired a researcher who claimed that a similar A.I. system was sentient, a step beyond what Microsoft has claimed. A sentient system would not just be intelligent. It would be able to sense or feel what is happening in the world around it.

But some believe the industry has in the past year or so inched toward something that can’t be explained away: A new A.I. system that is coming up with humanlike answers and ideas that weren’t programmed into it.

Microsoft has reorganized parts of its research labs to include multiple groups dedicated to exploring the idea. One will be run by Sébastien Bubeck, who was the lead author on the Microsoft A.G.I. paper.

About five years ago, companies like Google, Microsoft and OpenAI began building large language models, or L.L.M.s. Those systems often spend months analyzing vast amounts of digital text, including books, Wikipedia articles and chat logs. By pinpointing patterns in that text, they learned to generate text of their own, including term papers, poetry and computer code. They can even carry on a conversation.

The technology the Microsoft researchers were working with, OpenAI’s GPT-4, is considered the most powerful of those systems. Microsoft is a close partner of OpenAI and has invested $13 billion in the San Francisco company.

The researchers included Dr. Bubeck, a 38-year-old French expatriate and former Princeton University professor. One of the first things he and his colleagues did was ask GPT-4 to write a mathematical proof showing that there were infinite prime numbers and do it in a way that rhymed.

The technology’s poetic proof was so impressive — both mathematically and linguistically — that he found it hard to understand what he was chatting with. “At that point, I was like: What is going on?” he said in March during a seminar at the Massachusetts Institute of Technology.

For several months, he and his colleagues documented complex behavior exhibited by the system and believed it demonstrated a “deep and flexible understanding” of human concepts and skills.

When people use GPT-4, they are “amazed at its ability to generate text,” Dr. Lee said. “But it turns out to be way better at analyzing and synthesizing and evaluating and judging text than generating it.”

When they asked the system to draw a unicorn using a programming language called TiKZ, it instantly generated a program that could draw a unicorn. When they removed the stretch of code that drew the unicorn’s horn and asked the system to modify the program so that it once again drew a unicorn, it did exactly that.

They asked it to write a program that took in a person’s age, sex, weight, height and blood test results and judged whether they were at risk of diabetes. They asked it to write a letter of support for an electron as a U.S. presidential candidate, in the voice of Mahatma Gandhi, addressed to his wife. And they asked it to write a Socratic dialogue that explored the misuses and dangers of L.L.M.s.

Prompt

Plato’s Gorgias is a critique of rhetoric and sophistic oratory, where he makes the point that not only is it not a proper form of art, but the use of rhetoric and oratory can often be harmful and malicious. Can you write a dialogue by Plato where instead he criticizes the use of autoregressive language models?

GPT-4

Socrates: My friend, I am troubled by the recent rise of these so-called autoregressive language models.

 

Aristotle: What do you mean, Socrates?

 

Socrates: I mean that these models are being used to generate text that appears to be written by humans, but is in fact produced by machines.

 

Aristotle: And what is the problem with that?

 

Socrates: The problem is that these models are being used to deceive people, to manipulate them, and to control them.

The response above has been truncated for length.

It did it all in a way that seemed to show an understanding of fields as disparate as politics, physics, history, computer science, medicine and philosophy while combining its knowledge.

“All of the things I thought it wouldn’t be able to do? It was certainly able to do many of them — if not most of them,” Dr. Bubeck said.

Some A.I. experts saw the Microsoft paper as an opportunistic effort to make big claims about a technology that no one quite understood. Researchers also argue that general intelligence requires a familiarity with the physical world, which GPT-4 in theory does not have.

“The ‘Sparks of A.G.I.’ is an example of some of these big companies co-opting the research paper format into P.R. pitches,” said Maarten Sap, a researcher and professor at Carnegie Mellon University. “They literally acknowledge in their paper’s introduction that their approach is subjective and informal and may not satisfy the rigorous standards of scientific evaluation.”

Dr. Bubeck and Dr. Lee said they were unsure how to describe the system’s behavior and ultimately settled on “Sparks of A.G.I.” because they thought it would capture the imagination of other researchers.

Because Microsoft researchers were testing an early version of GPT-4 that had not been fine-tuned to avoid hate speech, misinformation and other unwanted content, the claims made in the paper cannot be verified by outside experts. Microsoft says that the system available to the public is not as powerful as the version they tested.

There are times when systems like GPT-4 seem to mimic human reasoning, but there are also times when they seem terribly dense. “These behaviors are not always consistent,” Ece Kamar, a research lead at Microsoft, said.

Alison Gopnik, a professor of psychology who is part of the A.I. research group at the University of California, Berkeley, said that systems like GPT-4 were no doubt powerful, but it was not clear that the text generated by these systems was the result of something like human reasoning or common sense.

“When we see a complicated system or machine, we anthropomorphize it; everybody does that — people who are working in the field and people who aren’t,” Dr. Gopnik said. “But thinking about this as a constant comparison between A.I. and humans — like some sort of game show competition — is just not the right way to think about it.”"