Täällä seurataan AI-ripuligeneraattoreiden maailmanvalloitusta

[pigra senlaborulo]

mikrobitti.fi
Tekoälytyökalujen pelätään suosivan miehiä ja ajavan naiset pois työelämästä
Justus Vento
~2 minutes

Neljäsosa kaikkien alojen työntekijöistä pelkää generatiivisen tekoälyn korvaavan työpaikoilla erityisesti naispuoliset työntekijät, kertoo Code First Girls -järjestön tuore tutkimus.

Yritysten pelätään kohdistavan tekoälysovelluksia etenkin niihin tehtäviin, joissa on tällä hetkellä paljon naispuolisia työntekijöitä. Asiasta kertoo ItPro.

Ilmiö koskee myös it-alaa. Se on perinteisesti ollut miesvaltainen ala, ja nyt se uhkaa yksipuolistua entisestään.

Jo kahdeksan kymmenestä työntekijästä käyttää järjestön mukaan ChatGPT:tä työssään. Nyt sen pelätään syrjäyttävän kokonaan ainakin osan työntekijöistä.

Code First Girlsin mukaan kielimallien kouluttamiseen käytettävä data on pahasti kallellaan miesten suuntaan, jolloin myös rekrytoinnit voivat painottua miespuolisten kandidaattien osaamiseen.

”Jos emme panosta koulutusmateriaalin monimuotoisuuteen, on riski, että kielimalleista tulee syrjiviä tiettyjä sukupuoli- tai muita vähemmistöjä kohtaan. Tästä syystä yritysten on tärkeää panostaa kielimalleja kouluttavien tiimien monimuotoisuuteen”, sanoo järjestön toimitusjohtaja Anna Brailsford.

Lausunto ei perustu ainoastaan oletukseen: viime vuonna julkaistun Kenan Instituten tutkimuksen mukaan kahdeksan kymmenestä naisesta työskentelee Yhdysvalloissa aloilla tai työtehtävissä, joita generatiiviset tekoälykielimallit uhkaavat korvata lähitulevaisuudessa. Miesten kohdalla lukema oli ”vain” kuusi kymmenestä.

Ilmiö ei ole uusi, tai ainoastaan uudempien kielimallien aiheuttama. Esimerkiksi Amazon joutui vuonna 2018 luopumaan rekrytoinneissa käytetystä tekoälystä, sillä se alkoi syrjiä naispuolisia hakijoita. Sen kouluttamiseen käytetty data oli peräisin pääosin miespuolisilta työntekijöiltä.

https://www.mikrobitti.fi/uutiset/mb/43 ... 56c58c23b5

[pigra senlaborulo]

Ohjelmistojätti Microsoft aikoo sijoittaa yhteensä 3,2 miljardia euroa Saksaan. Investointi ohjataan suurimmaksi osaksi tekoälykehitykseen.

Tarkoitus on kasvattaa yhtiön datakeskuksia ja kouluttaa tekoälyn kehittäjiä, sanoi yhtiön varatoimitusjohtaja ja hallituksen varapuheenjohtaja Brad Smith.

Smithin mukaan Saksa on teknologian eturintamassa, kun tarkastellaan tekoälyyn perustuvien sovellusten kehittämistä. Samalla maassa on kuitenkin pulaa osaajista, hän sanoi. Markkinat näyttävät kuitenkin lupaavilta muun muassa konepajateollisuudessa, autoteollisuudessa, lääketeollisuudessa ja rahoituspalveluissa.

Liittokansleri Olaf Scholz oli julkistamistilaisuudessa paikalla ja kiitteli sen osoittavan, miten kiinnostava ja houkutteleva sijoituskohde Saksa on.

https://www.hs.fi/talous/art-2000010231288.html

[pigra senlaborulo]

mikrobitti.fi
Monien suosikkiominaisuus Windowsissa sai väistyä Copilotin tieltä
Timo Tamminen
1–2 minutes

Windows 11 on saanut helmikuussa pakollisen päivityksen, joka kytkee ”näytä työpöytä” -ominaisuuden pois käytöstä. Tähän asti työpöydälle on voinut kurkata helposti viemällä hiiren osoittimen tehtäväpalkin oikeaan reunaan.

Vastaisuudessa ominaisuus on kuitenkin kytketty pois käytöstä, ja sen tilalla on Microsoftin tekoälyapuri Copilotin kuvake. Asiasta kirjoittaneella WindowsLatest-sivustolla oletetaan, että näytä työpöytä -painike sai väistyä, jotta Copilotille jää enemmän tilaa.

Copilot ei toistaiseksi ole saatavilla EU-alueella, joten Euroopassa asuvat Windows 11:n käyttäjät saanevat ainakin toistaiseksi pitää näytä työpöytä -ominaisuuden toiminnassa.

Näytä työpöytä -ominaisuus on ehtinyt kulkea Windowsin mukana pitkän matkan. Se on ollut monen mielestä kätevä tapa vilkaista nopeasti työpöydälle. Ominaisuutta on voinut hyödyntää kahdella eri tapaa: pikavilkaisulla ja sovellusikkunat pienentämällä.

Microsoft satsaa isosti uuteen tekoälyapuriinsa. Odotettavissa onkin, että Windowsin käyttäjä tulee törmäämään Copilotiin väkisinkin myös jatkossa.

https://www.mikrobitti.fi/uutiset/mb/29 ... e7978742fd

[Aloysius Kärppä]

https://www.cnbc.com/2024/02/14/nvidia- ... -firm.html

Nvidia passes Alphabet in market cap and is now the third most valuable U.S. company

KEY POINTS

The symbolic milestone is more confirmation that Nvidia has become a Wall Street darling on the back of elevated AI chips sales.

Nvidia is valued even more highly than the large software companies and cloud providers that developed many of the AI techniques that power the boom.

Nvidia is now the third largest U.S. company, only behind Apple and Microsoft.

[pigra senlaborulo]

mikrobitti.fi
Tekoälyt voivat povata datakeskuksille pahoja ongelmia – kasvava energiankulutus voi jopa ajaa konesaleja alas
Justus Vento
~2 minutes

Datakeskusten vaatima energiamäärä kasvaa koko ajan, ja generatiiviset tekoälysovellukset kasvattavat sitä vielä aiempaakin nopeammin. Datakeskukset voivat vaatia tulevaisuudessa 300–500 megawattia tehoa, mikä aiheuttaa entistä enemmän painetta sähköverkolle ja päivittäiselle toiminnalle, kertoo ItPro.

ItPron siteeraama JLL muistuttaa, että 1,3 biljoonan (1300 miljardia) arvoiseksi kymmenessä vuodessa paisuva tekoälybisnes vaatii datakeskuksilta myös huomattavasti erilaista infrastruktuuria kuin tavallisilta keskuksilta.

Niiden täytyy olla tiiviimmin pakattuja ja kestää suurempia tehoja. Tämä puolestaan kasvattaa huomattavasti jäähdytyslaitteistoja ja sitä kautta investoinnin kokoa.

Kasvavat tehot eivät ole ilouutinen sähköverkoille, jotka ovat suuressa osassa Eurooppaa yli 40 vuotta vanhoja. Niitä ei ole suunniteltu lähes täysikokoisen hiilivoimalan verran tehoa syöviä datakeskuksia ajatellen. Kulutuspiikit voisivat aiheuttaa jopa alasajoja.

Sähköverkkoinvestointien odotetaan nousevan maanosan laajuisesti satoihin miljardeihin euroihin.

JLL:n mukaan tekoälysovelluksiin erikoistuneissa datakeskuksissa on kuitenkin hyviäkin puolia. Tekoälysovellukset eivät tyypillisesti ole yhtä latenssiriippuvaisia kuin tavanomaisemmat sovellukset, jolloin niiden kehittäjillä on paremmat mahdollisuudet valita sijainti vakaan sähkömarkkinan perusteella keskeisen sijainnin sijaan.

Tutkijat tosin huomauttavat, että monessa paikassa riskinä on sähköpulan iskiessä automaattisesti aloitettavat alueelliset sähkökatkot. Ne ovat myrkkyä datakeskuksille, joten niiden sijainti täytyy valita tarkkaan.

https://www.mikrobitti.fi/uutiset/mb/db ... 78c62d88cb

[pigra senlaborulo]

bloomberg.com
Reddit Is Said to Sign AI Content Licensing Deal Ahead of IPO
Amy Or
2–3 minutes

Reddit Inc. has signed a contract allowing a company to train its artificial intelligence models on the social media platform’s content, according to people familiar with the matter, as it nears the potential launch of its long-awaited initial public offering.

The San Francisco-based firm told prospective investors in its IPO that it had signed the deal, worth about $60 million on an annualized basis, earlier this year, the people said. Reddit’s agreement with an unnamed large AI company could be a model for future contracts of a similar nature, one of the people said.

Reddit had more than $800 million in revenue last year, about a 20% increase over its 2022 figure, people familiar with the matter have said. The ability to profit from the AI wave sweeping the corporate world could help Reddit tap into investors’ enthusiasm for the technology, and boost its IPO.

The company has been advised to consider a valuation of at least $5 billion in an IPO, which could launch as soon as next month, Bloomberg News has reported.

Deliberations around the IPO are ongoing and details of both the listing and the AI deal could change, the people said, asking not to be identified as the information isn’t public. A representative for Reddit declined to comment.

AI companies hungry for data to train their programs on have been striking data licensing deals with content providers eager for new sources of revenue. OpenAI agreed a contract in December with German media giant company Axel Springer SE for AI training worth tens of millions of dollars, according to a Bloomberg News report.

The startup behind the ChatGPT text generation tool is also in talks with publishers including CNN, Fox Corp. and Time to license their work, Bloomberg News has reported, to feed its AI chatbots data to make their results more accurate, relevant and up to date.

About 16 banks, led by Morgan Stanley and Goldman Sachs Group Inc., are working on the IPO. JPMorgan Chase & Co., Bank of America Corp. and Citigroup Inc. are also among the banks on the deal.


https://www.bloomberg.com/news/articles ... ead-of-ipo

[pigra senlaborulo]

OpenAI collapses media reality with Sora, a photorealistic AI video generator
Hello, cultural singularity—soon, every video you see online could be completely fake.

Benj Edwards - 2/16/2024, 5:23 PM
Snapshots from three videos generated using OpenAI's Sora.
Enlarge / Snapshots from three videos generated using OpenAI's Sora.

On Thursday, OpenAI announced Sora, a text-to-video AI model that can generate 60-second-long photorealistic HD video from written descriptions. While it's only a research preview that we have not tested, it reportedly creates synthetic video (but not audio yet) at a fidelity and consistency greater than any text-to-video model available at the moment. It's also freaking people out.
Further Reading
AI-generated video of Will Smith eating spaghetti astounds with terrible beauty

"It was nice knowing you all. Please tell your grandchildren about my videos and the lengths we went to to actually record them," wrote Wall Street Journal tech reporter Joanna Stern on X.

"This could be the 'holy shit' moment of AI," wrote Tom Warren of The Verge.

"Every single one of these videos is AI-generated, and if this doesn't concern you at least a little bit, nothing will," tweeted YouTube tech journalist Marques Brownlee.

For future reference—since this type of panic will some day appear ridiculous—there's a generation of people who grew up believing that photorealistic video must be created by cameras. When video was faked (say, for Hollywood films), it took a lot of time, money, and effort to do so, and the results weren't perfect. That gave people a baseline level of comfort that what they were seeing remotely was likely to be true, or at least representative of some kind of underlying truth. Even when the kid jumped over the lava, there was at least a kid and a room.

The prompt that generated the video above: "A movie trailer featuring the adventures of the 30 year old space man wearing a red wool knitted motorcycle helmet, blue sky, salt desert, cinematic style, shot on 35mm film, vivid colors."

Technology like Sora pulls the rug out from under that kind of media frame of reference. Very soon, every photorealistic video you see online could be 100 percent false in every way. Moreover, every historical video you see could also be false. How we confront that as a society and work around it while maintaining trust in remote communications is far beyond the scope of this article, but I tried my hand at offering some solutions back in 2020, when all of the tech we're seeing now seemed like a distant fantasy to most people.

In that piece, I called the moment that truth and fiction in media become indistinguishable the "cultural singularity." It appears that OpenAI is on track to bring that prediction to pass a bit sooner than we expected.

Prompt: Reflections in the window of a train traveling through the Tokyo suburbs.

OpenAI has found that, like other AI models that use the transformer architecture, Sora scales with available compute. Given far more powerful computers behind the scenes, AI video fidelity could improve considerably over time. In other words, this is the "worst" AI-generated video is ever going to look. There's no synchronized sound yet, but that might be solved in future models.
How (we think) they pulled it off

AI video synthesis has progressed by leaps and bounds over the past two years. We first covered text-to-video models in September 2022 with Meta's Make-A-Video. A month later, Google showed off Imagen Video. And just 11 months ago, an AI-generated version of Will Smith eating spaghetti went viral. In May of last year, what was previously considered to be the front-runner in the text-to-video space, Runway Gen-2, helped craft a fake beer commercial full of twisted monstrosities, generated in two-second increments. In earlier video-generation models, people pop in and out of reality with ease, limbs flow together like pasta, and physics doesn't seem to matter.

Sora (which means "sky" in Japanese) appears to be something altogether different. It's high-resolution (1920x1080), can generate video with temporal consistency (maintaining the same subject over time) that lasts up to 60 seconds, and appears to follow text prompts with a great deal of fidelity. So, how did OpenAI pull it off?

OpenAI doesn't usually share insider technical details with the press, so we're left to speculate based on theories from experts and information given to the public.

OpenAI says that Sora is a diffusion model, much like DALL-E 3 and Stable Diffusion. It generates a video by starting off with noise and "gradually transforms it by removing the noise over many steps," the company explains. It "recognizes" objects and concepts listed in the written prompt and pulls them out of the noise, so to speak, until a coherent series of video frames emerge.
Further Reading
From toy to tool: DALL-E 3 is a wake-up call for visual artists—and the rest of us

Sora is capable of generating videos all at once from a text prompt, extending existing videos, or generating videos from still images. It achieves temporal consistency by giving the model "foresight" of many frames at once, as OpenAI calls it, solving the problem of ensuring a generated subject remains the same even if it falls out of view temporarily.

OpenAI represents video as collections of smaller groups of data called "patches," which the company says are similar to tokens (fragments of a word) in GPT-4. "By unifying how we represent data, we can train diffusion transformers on a wider range of visual data than was possible before, spanning different durations, resolutions, and aspect ratios," the company writes.

An important tool in OpenAI's bag of tricks is that its use of AI models is compounding. Earlier models are helping to create more complex ones. Sora follows prompts well because, like DALL-E 3, it utilizes synthetic captions that describe scenes in the training data generated by another AI model like GPT-4V. And the company is not stopping here. "Sora serves as a foundation for models that can understand and simulate the real world," OpenAI writes, "a capability we believe will be an important milestone for achieving AGI."

One question on many people's minds is what data OpenAI used to train Sora. OpenAI has not revealed its dataset, but based on what people are seeing in the results, it's possible OpenAI is using synthetic video data generated in a video game engine in addition to sources of real video (say, scraped from YouTube or licensed from stock video libraries). Nvidia's Dr. Jim Fan, who is a specialist in training AI with synthetic data, wrote on X, "I won't be surprised if Sora is trained on lots of synthetic data using Unreal Engine 5. It has to be!" Until confirmed by OpenAI, however, that's just speculation.
Sora as world simulator

Along with Sora, OpenAI released a corresponding technical document called "Video generation models as world simulators." That technical analysis merits a deeper dive than we have time or space for here, but how Sora models the world internally has computer scientists like Fan speculating about deeper things on X. "If you think OpenAI Sora is a creative toy like DALLE, ... think again. Sora is a data-driven physics engine," he wrote. "It is a simulation of many worlds, real or fantastical. The simulator learns intricate rendering, 'intuitive' physics, long-horizon reasoning, and semantic grounding, all by some denoising and gradient maths."

In the technical paper, OpenAI writes, "We find that video models exhibit a number of interesting emergent capabilities when trained at scale. These capabilities enable Sora to simulate some aspects of people, animals, and environments from the physical world. These properties emerge without any explicit inductive biases for 3D, objects, etc.—they are purely phenomena of scale."

OpenAI has also found that Sora can simulate Minecraft gameplay to some extent, bringing us one step closer to the potential of what might be called "neural rendering" in video games. Instead of rendering billions of polygons hand-crafted by artists, video game consoles of the future may generate interactive video streams using diffusion techniques in real time.

Sora is not perfect, however, and OpenAI notes Sora's deficiencies in its technical paper. "It does not accurately model the physics of many basic interactions, like glass shattering," the company writes. "Other interactions, like eating food, do not always yield correct changes in object state." OpenAI also lists "incoherencies that develop in long duration samples" and "spontaneous appearances of objects" as failures.

Here's an example of Sora when it doesn't do what you might expect to a glass sitting on a table.

There's also skepticism that tech like Sora may not be the universal solution to video generation. Computer scientist Grady Booch wrote, "I'm beginning to think that, while there will certainly be some economically- and creatively-interesting use cases, I see strong parallels to the domain of no code/low code efforts. In both those visual and programming domains, it is easy to produce splashy demos; it is easy to automate relative straightforward things. But nudging those systems to get the precise details you want? That’s another story."
Further Reading
AI image generation tech can now create life-wrecking deepfakes with ease

With a release like this, there are many dimensions of impact to consider, and we'll discuss those in future articles. Already, some are concerned about the implications for the film industry, the source of the training data, and the misinformation or disinformation that could come from being able to synthesize complex, high-resolution video on demand.

As a result, OpenAI says it is currently red-teaming Sora (putting it through adversarial testing) using "domain experts in areas like misinformation, hateful content, and bias" before it sees a public release. Even if OpenAI were to keep Sora locked in a vault forever, if history is any guide, open weights models will eventually catch up and similar technology will be available to all. Our main takeaway is this: If trusting video from anonymous sources on social media was a bad idea before, it's an even worse idea now.

https://arstechnica.com/information-tec ... generator/

linkin takana paljoa videota yms shits

[pigra senlaborulo]

tommonen videogenerointipaska varmaan syö vielä aivan saatanasti enemmän sähköä kuin kirjallinen ripulin generointi

[Ultra-lehti]

https://arstechnica.com/tech-policy/202 ... s-chatbot/

After months of resisting, Air Canada was forced to give a partial refund to a grieving passenger who was misled by an airline chatbot inaccurately explaining the airline's bereavement travel policy.

On the day Jake Moffatt's grandmother died, Moffat immediately visited Air Canada's website to book a flight from Vancouver to Toronto. Unsure of how Air Canada's bereavement rates worked, Moffatt asked Air Canada's chatbot to explain.

The chatbot provided inaccurate information, encouraging Moffatt to book a flight immediately and then request a refund within 90 days. In reality, Air Canada's policy explicitly stated that the airline will not provide refunds for bereavement travel after the flight is booked. Moffatt dutifully attempted to follow the chatbot's advice and request a refund but was shocked that the request was rejected.

[Ultra-lehti]

When Ars visited Air Canada's website on Friday, there appeared to be no chatbot support available, suggesting that Air Canada has disabled the chatbot.

[Spandau Mullet]

Vaimolle linkattiin eilen itsestään ripuligeneroitu "net worth" -sivu, jossa tätä tituleerataan näyttelijäksi ja kerrotaan nettovarallisuutensa olevan 12 miljoonaa dollaria. Onhan sillä tosielämässä yksi imdb-krediitti juuh. Nyt mietitään että missä ne rahat on

[Ilmestyskirjan soturi]

Vaimolle linkattiin eilen itsestään ripuligeneroitu "net worth" -sivu, jossa tätä tituleerataan näyttelijäksi ja kerrotaan nettovarallisuutensa olevan 12 miljoonaa dollaria. Onhan sillä tosielämässä yksi imdb-krediitti juuh. Nyt mietitään että missä ne rahat on

Onnea rikastumisen johdosta

[Spandau Mullet]

Vaimolle linkattiin eilen itsestään ripuligeneroitu "net worth" -sivu, jossa tätä tituleerataan näyttelijäksi ja kerrotaan nettovarallisuutensa olevan 12 miljoonaa dollaria. Onhan sillä tosielämässä yksi imdb-krediitti juuh. Nyt mietitään että missä ne rahat on

Onnea rikastumisen johdosta

Kiitos. Naurattaahaan se, mutta toisaalta vähän paska jos vaikka joku pahantahtoisempi ihminen googlailee että jaaha noilla olisi varmaan esim. kämpillä murtautumisen arvoisesti omaisuutta.

[Marxin Ryyppy]

Uusi hhot homma on AI-generoidut sienikirjat, voi jeesuksen vittu tämmöistä paskaa.

‘Life or Death:’ AI-Generated Mushroom Foraging Books Are All Over Amazon

Experts are worried that books produced by ChatGPT for sale on Amazon, which target beginner foragers, could end up killing someone.

A genre of AI-generated books on Amazon is scaring foragers and mycologists: cookbooks and identification guides for mushrooms aimed at beginners.

Amazon has an AI-generated books problem that’s been documented by journalists for months. Many of these books are obviously gibberish designed to make money. But experts say that AI-generated foraging books, specifically, could actually kill people if they eat the wrong mushroom because a guidebook written by an AI prompt said it was safe.

The New York Mycological Society (NYMS) warned on social media that the proliferation of AI-generated foraging books could “mean life or death.”

: PSA Alert!
: link in bio@Amazon and other retail outlets have been inundated with AI foraging and identification books.

Please only buy books of known authors and foragers, it can literally mean life or death. pic.twitter.com/FSqQLDhh42

— newyorkmyc (@newyorkmyc) August 27, 2023

tviitti kuvana::

Spoiler:

“There are hundreds of poisonous fungi in North America and several that are deadly,” Sigrid Jakob, president of the New York Mycological Society, told me in an email. “They can look similar to popular edible species. A poor description in a book can mislead someone to eat a poisonous mushroom.”

A quick scan of Amazon’s mushroom and foraging books revealed a bunch of books likely written by ChatGPT, but are sold without any indication that they’re AI-generated and are marketed as having been written by a human when they’re very likely not.

“Edwin J. Smith” is the author listed on two books—The Ultimate Mushroom Books Field Guide Of The Southwest: An essential field guide to foraging edible and non-edible mushrooms outdoors and indoors and Psilocybin Mushroom Book: Field Guide To Identification, Growing, and Microdosing Psilocybin Mushroom for Safe Use and Health Remedies—but doesn’t have any other books, or an online presence otherwise. The only Edwin J. Smith I could find was a Professor Emeritus of medicine at Indiana University from a staff list that’s more than a decade old.

I ran the descriptions of these books through AI-generated text detection tool ZeroGPT, which estimates how much of a selection of text might be written by an AI text generator like ChatGPT, and both were more than 85 percent written by AI. In the sample pages available on the Amazon listings, both follow a specific format: a short piece of anecdotal fiction about a hobbyist. The psilocybin mushroom book begins (which ZeroGPT determined is 94 percent AI-written):

“Stella had always been interested in the natural world and was fascinated by the many unknowns of the plant world. She had heard about the power of a special type of mushroom, known as psilocybin mushrooms, and decided to dedicate her time to learning more about them.”

And the Southwest field guide begins with:

“Tompson was a health enthusiast, always looking for new ways to improve his diet and overall wellness. He loved incorporating fresh vegetables and herbs into his meals and was particularly fond of mushrooms. However, he found that buying mushrooms at the grocery store was becoming increasingly expensive, and he longed for a more sustainable and cost-effective solution.”

ZeroGPT ranked those lines as 100 percent GPT.

Another sample page from The Ultimate Mushroom Books Field Guide Of The Southwest, which briefly notes that some mushrooms can be poisonous before pivoting to medicinal properties of mushrooms, was also likely generated by AI, according to ZeroGPT.


A SAMPLE PAGE FROM THE ULTIMATE MUSHROOM BOOKS FIELD GUIDE OF THE SOUTHWEST

Another book, WILD MUSHROOM COOKBOOK FOR BEGINNER: The complete guide on mushroom foraging and cooking with delicious recipes to enjoy your favorite, was written by “Mageret Lawrence,” who has published 19 titles—mostly how-to guides about diet and nutrition—on Amazon within the last two months. Her author page is suspiciously vague, citing a nutrition background and research experience, but no specifics as to which institutions she worked at or earned degrees from. The description for this cookbook may be at least partially generated by ChatGPT, according to ZeroGPT.


THREE BOOKS SUSPECTED OF BEING AI-GENERATED, ACCORDING TO THE DETECTION TOOLS 404 MEDIA USED.

[lopputeksti spoilerin takana]

Spoiler:

Another potentially AI-generated author, “Dr. Kimberly Thorpe,” has written 37 nutrition and health books on Amazon, including one titled WILD MUSHROOM COOKBOOK: A Beginner's guide to learning the basics of cooking with wild mushrooms for health and flavor, complete with easy-to-follow recipes!. All of Thorpe’s dozens of books have been published in the last several months. They are all in the same style as the above examples, and the GPT detector said it’s 100 percent AI. Unlike the others, Thorpe lists the university where she got her Ph.D, and mentions having written for the Huffington Post and the New York Times, but I couldn’t find her in any of these places. A spokesperson for UC Davis told me that they checked the registrar and do not have any alumni by that name in their records.



After 404 Media reached out for comment and sent the company links to these suspected AI books, Amazon deleted The Ultimate Mushroom Books Field Guide Of The Southwest, Psilocybin Mushroom Book, and WILD MUSHROOM COOKBOOK FOR BEGINNER.

“Amazon is constantly evaluating emerging technologies and is committed to providing the best possible shopping, reading, and publishing experience for our authors and customers,” Amazon spokesperson Ashley Vanicek told 404 Media. “All publishers in the store must adhere to our content guidelines, regardless of how the content was created. We invest significant time and resources to ensure our guidelines are followed, and remove books that do not adhere to these guidelines. We’re committed to providing a safe shopping and reading experience for our customers and we take matters like this seriously.”

False Morels and Death Caps are two species found in the American Southwest that look a lot like their edible, non-poisonous counterparts and can kill you within hours. Foraging safely for mushrooms can require deep fact checking, curating multiple sources of information, and personal experience with the organism, Jakob said—none of which ChatGPT has the ability to do. “Human written books can take years to research and write,” Jakob said.

The differences between a mushroom that will kill you and one that’ll work well on a pizza can be difficult even for humans to spot. Elan Trybuch, secretary for the New York Mycological Society, told me that it can require first hand interaction with the shrooms to determine if they’re deadly.

“AI does not know the subtle differences between a mushroom that is poisonous such as the Death cap (Amanita bisporigera) vs one that is not, Wood Mushroom (Agaricus silvicola),” Trybuch said. “NYMS members have even found people foraging in Harriman State Park, who before AI books, were foraging for the Amanita bisporigera. Luckily we saw them and were able to explain to them the differences. One way we do so on our walks, is to use KOH (Potassium Hydroxidie) to see the color change on the cap itself.”

]Destroying Angels aka Amanita bisporigera are easily mistaken as edible. See pic below: #amanita #deadly #fungi #mushroom #foraging pic.twitter.com/huIuPOWZFg

— newyorkmyc (@newyorkmyc) September 25, 2021

tviitti kuvana:

Spoiler:

Ben Colman, co-founder and CEO of AI manipulation detection tool Reality Defender, told 404 Media that the author images are most likely generated by an AI. “The two author images are almost certainly GAN-based creations, as displayed in our results. GANs are one of many thousands of methods used to generate deepfakes and AI-generated images,” he said.



When run through a Hugging Face open-source AI image detector, the author photos scored as highly likely to be machine generated. Lawrence’s photo contains visible artifacts around the earlobes and hair that are consistent with AI-generated images like the ones found on Linkedin scam profiles.

Looking for authors’ activity elsewhere online is one of the ways mycologists and experienced foragers suggest finding reliable information. Rebecca Lexa, who writes books on the natural world herself, and shares naturalist information like identification tips, sounded the alarm on Tumblr about AI-generated foraging books, and used an Amazon-listed author named “Jason Cones” as an example: Cones has 26 books on Amazon (including one aimed at foraging beginners) but no other online presence. “If you search for him online, the only page for an author named Jason Cones is the Amazon author page--no website, no social media, no interviews, nada,” Lexa wrote. “Even a brand new author will at least have something other than their Amazon page, and they'll mention experience, credentials, other biographical info.” Cones’ author bio, like Lawrence’s, is vague about his experience.

One of the beautiful things about the foraging community is its openness to beginners: groups like NYC Mycology Society take new foragers around cities and parks to show them firsthand how to safely forage from public spaces, and amateur foragers contribute valuable research and science to mycology. But when it’s flooded with ChatGPT trash, it’ll only get harder to sift through the garbage to find morsels of quality, human-made info for beginners.

[pigra senlaborulo]

theguardian.com
OpenAI boss Sam Altman wants $7tn. For all our sakes, pray he doesn’t get it | John Naughton
John Naughton
5–6 minutes

Once upon a time, nobody outside tech circles had heard of Sam Altman. But then his company, OpenAI, launched ChatGPT, and suddenly he was everywhere – touring the world, giving interviews to gushing journalists, granting audiences to awestruck politicians etc. Whiplash-thin, with a charmingly wide-eyed baby face, he instantly became the acceptable face of digital capitalism.

Then the OpenAI board abruptly fired him, apparently on the grounds that he had not been, er, entirely candid with them. When Satya Nadella, Microsoft’s CEO (who had invested $13bn in OpenAI), heard about it, though, he was mightily pissed off. And in no time all, Altman was unsacked and reinstated in the OpenAI driving seat. And the world was transfixed by the drama of it all. Which only goes to show that appearances can be deceptive.

If the world had read Tad Friend’s profile of Altman, which appeared in the New Yorker in 2016, it might have been less overawed. “I have narrow interests in technology,” he told Friend. “I have no patience for things I’m not interested in: parties, most people. When someone examines a photo and says, ‘Oh, he’s feeling this and this and this,’ all these subtle emotions, I look on with alien intrigue.” Altman’s great strengths, concluded Friend, “are clarity of thought and an intuitive grasp of complex systems. His great weakness is his utter lack of interest in ineffective people, which unfortunately includes most of us. I found his assiduousness alarming at first, then gradually endearing.”

Two recent events suggest that it might be time to dial down the “endearing” bit. The first was the revelation that OpenAI was rowing back on its previous aversion to “military and warfare” use of its technology. The second was the announcement that Altman was wooing the United Arab Emirates for up to $7tn (£5.6tn) for the business of chips and AI. To put Altman’s aspirations in context, the sum he’s seeking to raise is just under a third of US GDP and pretty close to its £6.3tn federal budget for 2022. And, on a historical note, it’s $3tn more than the $4tn (adjusted for inflation) that the US spent on the second world war.

So what would $7tn get you? Well, as the Register helpfully points out, it’s enough cash “to gobble up Nvidia, TSMC, Broadcom, ASML, Samsung, AMD, Intel, Qualcomm, and every other chipmaker, designer, intellectual property holder, and equipment vendor of consequence in their entirety – and still have trillions left over”. But Altman doesn’t just aim to become the John D Rockefeller of our times – owning everything. He wants to make things – specifically the GPUs (graphics processing units) that machine-learning systems require. That means building semiconductor fabrication units (fabs). These cost about $20bn each and take four or more years to commission and become productive. They also require very highly skilled staff – which the US semiconductor industry is short of by about 70,000.

In addition, these plants are huge consumers of water, in a world that is rapidly running short of it. But Altman would have enough dosh to build 350 of the monsters. I could go on, but you get the message. This aspiration seems crazy. And yet the Silicon Valley crowd think he’s a genius. So what’s going on?

The answer is that most of them belong to the church of technocracy, of which Altman is a charismatic member. Devout members of this sect believe that the world is terminally screwed-up, and that the only way to fix it is with tech. They are ecstatic about AI because finally a technology has arrived that apparently could fix everything – economic growth, healthcare, productivity, education, even the climate crisis. Strangely, though, warfare seems to be missing from the list.

The only difficulty is that this magical technology needs unconscionable quantities of data and computational power. Our future, apparently, depends on infinite amounts of what the industry now calls “compute”, and Altman is lauded because only he has had the courage to say out loud how much of it is needed in order to save civilisation.

He is deeply conscious of the responsibility he carries. “Democracy only works in a growing economy,” he told Friend in 2016. “Without a return to economic growth, the democratic experiment will fail.” If it does, though, Altman will be ready. In a discussion about aggressive AI and nations fighting with nuclear weapons over scarce resources, he said: “I try not to think about it too much. But I have guns, gold, potassium iodide, antibiotics, batteries, water, gas masks from the Israel Defense Forces, and a big patch of land in Big Sur I can fly to.” Nice to know that that $7tn will be in safe hands.
What I’ve been reading

Tech groupthink
Adrienne LaFrance has written a terrific essay in the Atlantic on the underpinning techno-authoritarian ideology of Silicon Valley.

Dead hand
Robert Hutton has a Swiftian take on Britain’s zombie Conservatives 14 years on in the Critic magazine.

Sing it proud
The former US employment secretary, Robert Reich, has written a thoughtful Substack post titled “Why I preach to the choir”.

https://www.theguardian.com/commentisfr ... snt-get-it