Oracle cofounder Larry Ellison envisions AI as the backbone of a new era of mass surveillance, positioning Oracle as a key player in AI infrastructure through its unique networking architecture and partnerships with AWS and Microsoft. The Register reports: Ellison made the comments near the end of an hour-long chat at the Oracle financial analyst meeting last week during a question and answer session in which he painted Oracle as the AI infrastructure player to beat in light of its recent deals with AWS and Microsoft. Many companies, Ellison touted, build AI models at Oracle because of its "unique networking architecture," which dates back to the database era.

"AI is hot, and databases are not," he said, making Oracle's part of the puzzle less sexy, but no less important, at least according to the man himself - AI systems have to have well-organized data, or else they won't be that valuable. The fact that some of the biggest names in cloud computing (and Elon Musk's Grok) have turned to Oracle to run their AI infrastructure means it's clear that Oracle is doing something right, claimed now-CTO Ellison. "If Elon and Satya [Nadella] want to pick us, that's a good sign - we have tech that's valuable and differentiated," Ellison said, adding: One of the ideal uses of that differentiated offering? Maximizing AI's pubic security capabilities.

"The police will be on their best behavior because we're constantly watching and recording everything that's going on," Ellison told analysts. He described police body cameras that were constantly on, with no ability for officers to disable the feed to Oracle. Even requesting privacy for a bathroom break or a meal only meant sections of recording would require a subpoena to view - not that the video feed was ever stopped. AI would be trained to monitor officer feeds for anything untoward, which Ellison said could prevent abuse of police power and save lives. [...] "Citizens will be on their best behavior because we're constantly recording and reporting," Ellison added, though it's not clear what he sees as the source of those recordings - police body cams or publicly placed security cameras. "There are so many opportunities to exploit AI," he said.

Microsoft has abandoned plans to overhaul its Edge browser interface, scrapping the design choice unveiled in February 2023. The redesign -- featuring a sleeker look with rounded tab buttons and increased blur effects -- aimed to give Edge a distinct identity as the company pushed into AI services. The new design never officially launched and the company has no intention to launch it later, according to Microsoft-focused news outlet Windows Central.

A Microsoft spokesperson confirmed to Windows Central that the company is moving away from the rounded tabs concept. Some elements of the redesign will remain, including webpage borders and a repositioned user button, but the majority of the proposed changes have been shelved. The decision marks a retreat from Microsoft's efforts to visually differentiate Edge from Google Chrome and align it with Windows 11's design language.

"A two-day AI Labor Summit between AFL-CIO leaders and Microsoft executives this week reflects the tech giant's revamped approach to unions," writes GeekWire, "which includes a pledge by the company to incorporate feedback from labor unions and their members into the development of artificial intelligence."

But just two days later, "Microsoft Gaming CEO Phil Spencer announced it was game over for the jobs of another 650 Microsoft staffers (on top of an earlier 1,900 employee staff reduction)," writes long-time Slashdot reader theodp, "cuts that Spencer made clear were related to Microsoft's $69B acquisition of Activision Blizzard in 2023." Interestingly, Microsoft's Smith in October 2023 affirmed a "groundbreaking neutrality agreement" with the Communications Workers of America union (CWA) — designed to go into effect if Microsoft was successful in its acquisition of Activision Blizzard — in which Microsoft acknowledged the rights of its employees to unionize and pledged to work constructively with any who did. At the same time, Microsoft made it clear that it hoped its employees wouldn't feel the need to form or join unions, saying they would "never need to organize to have a dialogue with Microsoft's leaders."

In July 2023, the AFL-CIO applauded Microsoft's Activision Blizzard acquisition and the Microsoft-CWA agreement, which AFL-CIO union federation president Liz Shuler said "sets a new standard for respecting workers' rights in the video game industry and the larger technology sector." And in December 2023, Shuler thanked Smith for Microsoft's "absolutely historic partnership" on AI and the Future of the Workforce, which Shuler suggested "can be mutually beneficial for workers, for businesses, and for our country as a whole."
Thursday the CWA union issued critical remarks about the layoffs at Microsoft Gaming (which were later retweeted by the @AFLCIO Twitter account).

"While we would hope that a company like Microsoft with $88 billion in profits last year could achieve 'long-term success' without destroying the livelihoods of 650 of our colleagues, heartless layoffs like these have become all too common."

The Washingon Post reports that a regulatory dispute in Ohio may help answer a big question about America's power grid: who will pay for the huge upgrades needed to meet soaring energy demand "from the data centers powering the modern internet and artificial intelligence revolution?" Google, Amazon, Microsoft and Meta are fighting a proposal by an Ohio power company to significantly increase the upfront energy costs they'll pay for their data centers, a move the companies dubbed "unfair" and "discriminatory" in documents filed with Ohio's Public Utility Commission last month. American Electric Power Ohio said in filings that the tariff increase was needed to prevent new infrastructure costs from being passed on to other customers such as households and businesses if the tech industry should fail to follow through on its ambitious, energy-intensive plans. The case could set a national precedent that helps determine whether and how other states force tech firms to be accountable for the costs of their growing energy consumption... The energy demands of data centers have created similar concerns in other hot spots such as Northern Virginia, Atlanta and Maricopa County, Arizona, leaving experts concerned that the U.S. power grid may not be capable of dealing with the combined needs of the green energy transition and the computing boom that artificial intelligence companies say is coming...

Energy customers must sometimes make a monthly payment to a utility that is a percentage of the maximum amount of electricity they predict that they could need. In Ohio, data center companies had agreed to pay 60 percent of the projected amount. But in May, the power company proposed a new, 10-year fee structure raising the charges to 90 percent of the expected load, even if they don't end up using that much. The major tech companies — all of whom are increasing spending on data center infrastructure to compete in AI — strenuously opposed the proposed contract in documents filed last month... According to testimony from AEP Ohio Vice President Lisa Kelso, there are 50 pending requests from data center customers seeking electric service at more than 90 sites, a potential 30,000 megawatts of additional load — enough to power more than 20 million households. That additional demand would more than triple the utility's previous peak load in 2023, she said. Between 2020 and 2024, the data center energy load in central Ohio increased sixfold, from 100 to 600 megawatts, her testimony reads. By 2030, that amount will reach 5,000 megawatts, according to the utility's signed agreements, she testified...

Meeting that demand will require AEP Ohio to build new transmission lines, an expensive and time-consuming process... Chief among the power company's concerns, according to the documents, is what will happen if it invests billions of dollars into new grid infrastructure only for the data centers to leave for greener pastures, or for the AI bubble to burst and the facilities to need much less power than initially projected. If the power company spends big on new infrastructure but the power demand it was built to serve doesn't materialize, other customers — including business and residential payers — will be stuck with the bill, the utility said... AEP Ohio's testimony in the case also questions whether data centers bring as much to local communities as factories or other high-energy-load businesses. Since 2019, non-data center businesses have created approximately 25 jobs for every megawatt of power requested, while data centers have created less than one job per megawatt, according to Kelso's testimony.

The tech companies rejected this criticism, saying the number of jobs they create is not relevant to how much power they have a right to purchase, and highlighted their other contributions to local economies... Amazon said in filings that it pays fees as high as 75 percent of projected demand in some states but that Ohio's proposal to bill it 90 percent goes too far.
"Should the Ohio tariff be approved, Microsoft and Google both threatened in their testimony to leave Ohio." (Although at the same time, "pressure on the electric grid is mounting all over the country...")

And the article points out that on Thursday, "the White House announced measures intended to speed up data center construction for AI projects, including by accelerating permitting."

Cybersecurity giant Fortinet has confirmed it suffered a data breach after a threat actor claimed to steal 440GB of files from the company's Microsoft Sharepoint server. From a report: Fortinet is one of the largest cybersecurity companies in the world, selling secure networking products like firewalls, routers, and VPN devices. The company also offers SIEM, network management, and EDR/XDR solutions, as well as consulting services.

Early this morning, a threat actor posted to a hacking forum that they had stolen 440GB of data from Fortinet's Azure Sharepoint instance. The threat actor then shared credentials to an alleged S3 bucket where the stolen data is stored for other threat actors to download. The threat actor, known as "Fortibitch," claims to have tried to extort Fortinet into paying a ransom, likely to prevent the publishing of data, but the company refused to pay. In response to our questions about incident, Fortinet confirmed that customer data was stolen from a "third-party cloud-based shared file drive."

An anonymous reader shares a report: Microsoft executives have been thinking lately about the end of the world. In a white paper published late last year, Brad Smith, the company's vice chair and president, and Melanie Nakagawa, its chief sustainability officer, described a "planetary crisis" that AI could help solve. Imagine an AI-assisted tool that helps reduce food waste, to name one example from the document, or some future technology that could "expedite decarbonization" by using AI to invent new designs for green tech.

But as Microsoft attempts to buoy its reputation as an AI leader in climate innovation, the company is also selling its AI to fossil-fuel companies. Hundreds of pages of internal documents I've obtained, plus interviews I've conducted over the past year with 15 current and former employees and executives, show that the tech giant has sought to market the technology to companies such as ExxonMobil and Chevron as a powerful tool for finding and developing new oil and gas reserves and maximizing their production -- all while publicly committing to dramatically reduce emissions.

Although tech companies have long done business with the fossil-fuel industry, Microsoft's case is notable. It demonstrates how the AI boom contributes to one of the most pressing issues facing our planet today -- despite the fact that the technology is often lauded for its supposed potential to improve our world, as when Sam Altman testified to Congress that it could address issues such as "climate change and curing cancer." These deals also show how Microsoft can use the vagaries of AI to talk out of both sides of its mouth, courting the fossil-fuel industry while asserting its environmental bona fides. (Many of the documents I viewed have been submitted to the Securities and Exchange Commission as part of a whistleblower complaint alleging that the company has omitted from public disclosures "the serious climate and environmental harms caused by the technology it provides to the fossil fuel industry," arguing that the information is of material and financial importance to investors.

In a statement today, the White House said it has received commitments from several AI companies to curb the creation and distribution of deepfake porn, also known as image-based sexual abuse material. Engadget reports: The participating businesses have laid out the steps they are taking to prevent their platforms from being used to generate non-consensual intimate images (NCII) of adults and child sexual abuse material (CSAM). Specifically, Adobe, Anthropic, Cohere, Common Crawl, Microsoft and OpenAI said they'll be: "responsibly sourcing their datasets and safeguarding them from image-based sexual abuse."

All of the aforementioned except Common Crawl also agreed they'd be: "incorporating feedback loops and iterative stress-testing strategies in their development processes, to guard against AI models outputting image-based sexual abuse" and "removing nude images from AI training datasets" when appropriate. [...] The notable absences from today's White House release are Apple, Amazon, Google and Meta.

Microsoft announced plans to modify Windows, enabling security vendors like CrowdStrike to operate outside the operating system's kernel. The move follows the July incident where a faulty CrowdStrike update caused widespread system failures. From a report: Microsoft says it has now "discussed the requirements and key challenges in creating a new platform which can meet the needs of security vendors" with partners like CrowdStrike, Broadcom, Sophos, and Trend Micro.

[...] While Microsoft isn't directly saying it's going to close off access to the Windows kernel, it's clearly at the early stages of designing a security platform that can eventually move CrowdStrike and others out of the kernel. Microsoft last tried to close off access to the Windows kernel in Windows Vista in 2006, but it was met with pushback from cybersecurity vendors and regulators.

Microsoft is laying off a further 650 staff from its gaming business, according to a memo sent by Xbox chief Phil Spencer to staff today, September 12. From a report: In the memo, Spencer said the roles affect mostly corporate and support functions, and were made "to organize our business for long term success." He clarified that no games, devices or experiences are being canceled and no studios are being closed as part of these cuts. These latest layoffs mean Microsoft has let go of 2,550 staff from its gaming business since acquiring Activision Blizzard for $69 billion in 2023. In his memo, below, Spencer makes it clear that the cuts are related to the acquisition.

An anonymous reader quotes a report from Ars Technica: Microsoft has updated a key cryptographic library with two new encryption algorithms designed to withstand attacks from quantum computers. The updates were made last week to SymCrypt, a core cryptographic code library for handing cryptographic functions in Windows and Linux. The library, started in 2006, provides operations and algorithms developers can use to safely implement secure encryption, decryption, signing, verification, hashing, and key exchange in the apps they create. The library supports federal certification requirements for cryptographic modules used in some governmental environments. Despite the name, SymCrypt supports both symmetric and asymmetric algorithms. It's the main cryptographic library Microsoft uses in products and services including Azure, Microsoft 365, all supported versions of Windows, Azure Stack HCI, and Azure Linux. The library provides cryptographic security used in email security, cloud storage, web browsing, remote access, and device management. Microsoft documented the update in a post on Monday. The updates are the first steps in implementing a massive overhaul of encryption protocols that incorporate a new set of algorithms that aren't vulnerable to attacks from quantum computers. [...]

The first new algorithm Microsoft added to SymCrypt is called ML-KEM. Previously known as CRYSTALS-Kyber, ML-KEM is one of three post-quantum standards formalized last month by the National Institute of Standards and Technology (NIST). The KEM in the new name is short for key encapsulation. KEMs can be used by two parties to negotiate a shared secret over a public channel. Shared secrets generated by a KEM can then be used with symmetric-key cryptographic operations, which aren't vulnerable to Shor's algorithm when the keys are of a sufficient size. [...] The other algorithm added to SymCrypt is the NIST-recommended XMSS. Short for eXtended Merkle Signature Scheme, it's based on "stateful hash-based signature schemes." These algorithms are useful in very specific contexts such as firmware signing, but are not suitable for more general uses. Monday's post said Microsoft will add additional post-quantum algorithms to SymCrypt in the coming months. They are ML-DSA, a lattice-based digital signature scheme, previously called Dilithium, and SLH-DSA, a stateless hash-based signature scheme previously called SPHINCS+. Both became NIST standards last month and are formally referred to as FIPS 204 and FIPS 205. In Monday's post, Microsoft Principal Product Manager Lead Aabha Thipsay wrote: "PQC algorithms offer a promising solution for the future of cryptography, but they also come with some trade-offs. For example, these typically require larger key sizes, longer computation times, and more bandwidth than classical algorithms. Therefore, implementing PQC in real-world applications requires careful optimization and integration with existing systems and standards."

OpenAI is in talks to raise $6.5 billion from investors at a valuation of $150 billion, Bloomberg reported Wednesday, citing people familiar with the situation. From the report: The new valuation, a figure that doesn't include the money being raised, is significantly higher than the $86 billion valuation from the company's tender offer earlier this year, and cements its place as one of the most valuable startups in the world. At the same time, OpenAI is also in talks to raise $5 billion in debt from banks in the form of a revolving credit facility, said one of the people, all of whom asked not to be identified discussing private information.

The funding round is slated to be led by Thrive Capital, Bloomberg previously reported. Thrive declined to comment on the latest valuation. Microsoft, the company's largest investor, is also set to participate, and Apple and Nvidia, have been in talks about investing.

A security researcher has exposed a critical vulnerability in the WHOIS system. Benjamin Harris, CEO of watchTowr, gained unprecedented access by registering an expired domain once used for .mobi's authoritative WHOIS server. His rogue server received millions of queries from thousands of systems, including government agencies, certificate authorities, and major tech companies. ArsTechnica adds: The humor aside, the rogue WHOIS server gave him powers he never should have had. One of the greatest was the ability to dictate the email address certificate authority GlobalSign used to determine if a party applying for a TLS certificate was the rightful owner of the domain name the certificate would apply to. Like the vast majority of its competitors, GlobalSign uses an automated process. An application for example.com, for instance, will prompt the certificate authority to send an email to the administrative email address listed in the authoritative WHOIS for that domain. If the party on the other end clicks a link, the certificate is automatically approved. When Harris generated a certificate signing request for microsoft.mobi, he promptly received an email from GlobalSign. The email gave him the option of receiving a verification link at whois@watchtowr.com. For ethical reasons, he stopped the experiment at this point. The vulnerability stems from outdated WHOIS client configurations, which underscores systemic weaknesses in internet infrastructure management.

An anonymous reader quotes a report from Data Center Dynamics: Google has signed a $10 million deal to pull 100,000 tons of carbon dioxide out of the air. The company will buy direct air capture (DAC) credits from startup Holocene, to be delivered in the early 2030s. The deal is the lowest price on record for DAC, at $100 per ton -- a price the Department of Energy previously said was needed to make carbon capture mainstream. Google will provide the funds up front, but there is no guarantee that Holocene will hit that goal. Running Tide, a carbon removal company that Microsoft paid to capture 12,000 tons of CO2 in 2023, shut down in 2024. The $100 price was also made possible thanks to the US government's 45Q tax credit, which provides DAC suppliers $180 per ton of carbon removed.

Holocene passes air through a waterfall with an amino acid added to it which binds CO2. This is then mixed with guanidine to form a solid crystal mass. Next, the amino acid is sent back to the beginning of the loop, while the solid is lightly heated to release pure CO2 -- which can then be stored. The company plans to capture and store 100,000 tons of CO2 by the early 2030s. "The structure of this partnership -- providing immediate funding to achieve an ambitious but important price in the medium term -- is just one way to support carbon removal as it scales," Randy Spock, carbon credits and removals lead, said.

wiredmikey shares a report from SecurityWeek: Microsoft on Tuesday raised an alarm for in-the-wild exploitation of a critical flaw in Windows Update, warning that attackers are rolling back security fixes on certain versions of its flagship operating system. The Windows flaw, tagged as CVE-2024-43491 and marked as actively exploited, is rated critical and carries a CVSS severity score of 9.8/10. Redmond's documentation of the bug suggests a downgrade-type attack similar to the 'Windows Downdate' issue discussed at this year's Black Hat conference. Microsoft's bulletin reads: "Microsoft is aware of a vulnerability in Servicing Stack that has rolled back the fixes for some vulnerabilities affecting Optional Components on Windows 10, version 1507 (initial version released July 2015). This means that an attacker could exploit these previously mitigated vulnerabilities on Windows 10, version 1507 (Windows 10 Enterprise 2015 LTSB and Windows 10 IoT Enterprise 2015 LTSB) systems that have installed the Windows security update released on March 12, 2024 -- KB5035858 (OS Build 10240.20526) or other updates released until August 2024. All later versions of Windows 10 are not impacted by this vulnerability."

To protect against this exploit, Microsoft says Windows users should install this month's Servicing stack update (SSU KB5043936) and the September 2024 Windows security update (KB5043083), in that order.

Earlier today, Sony unveiled the $699.99 PlayStation 5 Pro -- a mid-generation upgrade model for the PlayStation 5 that requires a separate $79.99 disc drive if you want to play your physical games. As The Verge's Jay Peters writes, the announcement "may have marked the beginning of the end for game console disc drives." From the report: Microsoft is clearly eyeing the discless direction with Xbox as well. The more affordable Xbox Series S can't play discs, and there's a discless Series X in white that's set to launch later this year. Last year's giant Xbox leak revealed a cylindrical, "adorably all digital" Xbox Series X redesign too. That hasn't been announced as an official product, but it shows a disc-free future is on Microsoft's mind. It seems likely that Sony and Microsoft are testing the waters for going all-digital for the PlayStation 6 and the next generation Xbox -- or at least offering disc drives separately. It's like Apple removing the disc drive all over again.

But this time it's not just the people making the devices. Retailers are stepping back from physical media, too. Redbox is toast. Best Buy said last year that it would stop carrying physical movies, and Target recently confirmed that it would be all but ditching DVDs in its physical stores. I know movies aren't video games, but it doesn't seem like a big leap that brick-and-mortar stores might stop carrying physical video games down the line; UK retailer GAME has already ended video game trade-ins.

Microsoft today announced significant strides in its Azure Quantum Cloud service, including the demonstration of logical operations using the largest number of error-corrected qubits ever achieved. This progress brings the industry closer to building reliable quantum computers capable of solving complex problems beyond the reach of classical systems, the company said.

In a significant partnership, Microsoft is collaborating with Atom Computing to integrate their neutral-atom hardware into Azure Quantum. Atom Computing has already shown promise with hardware exceeding 1,000 qubits. Key to Microsoft's advancements is the implementation of the "tesseract code" error correction scheme on Quantinuum's trapped-ion quantum hardware. This led to a 22-fold reduction in error rates, a critical step towards reliable quantum computations. Microsoft is also committed to simplifying quantum programming. Azure's Q# language will now automatically handle complex error correction, making quantum development more accessible.

Oracle missed the tech industry's move to cloud computing last decade and ended up an also-ran. Now the AI boom has given it another shot. WSJ: The 47-year-old company that made its name on relational database software has emerged as an attractive cloud-computing provider for AI developers such as OpenAI, sending its long-stagnant stock to new heights. Oracle shares are up 34% since January, well outpacing the Nasdaq's 14% rise and those of bigger competitors Microsoft, Amazon.com and Google.

It is a surprising revitalization for a company many in the tech industry had dismissed as a dinosaur of a bygone, precloud era. Oracle appears to be successfully making a case to investors that it has become a strong fourth-place player in a cloud market surging thanks to AI. Its lateness to the game may have played to its advantage, as a number of its 162 data centers were built in recent years and are designed for the development of AI models, known as training.

In addition, Oracle isn't developing its own large AI models that compete with potential clients. The company is considered such a neutral and unthreatening player that it now has partnerships with Microsoft, Google and Amazon, all of which let Oracle's databases run in their clouds. Microsoft is also running its Bing AI chatbot on Oracle's servers.

CrowdStrike CFO Burt Podbere says the cybersecurity firm has not faced lawsuits over July's global IT outage. Speaking at a conference, Podbere emphasized efforts to shift customer focus from legal threats to business discussions. The Register: There were dark rumblings from Delta Air Lines last month, for example, threatening litigation over alleged gross negligence. At the time, CrowdStrike reiterated its apologies, saying: "Public posturing about potentially bringing a meritless lawsuit against CrowdStrike as a long-time partner is not constructive to any party." During his time at the Citi conference, Podbere admitted: "We don't know how it's all going to shake out.

"Everything we're doing and trying to do is take the legal discussion away from our interaction with customers and move it to the business discussion. "And as time goes on, that does get easier because we're moving further away from the Sun, right? And that's how we think about it."

Wired interviews America's foreign policy chief, Secretary of State Antony Blinken, about U.S. digital polices, starting with a new "cybersecurity bureau" created in 2022 (which Wired previously reported includes "a crash course in cybersecurity, telecommunications, privacy, surveillance, and other digital issues.") Look, what I've seen since coming back to the State Department three and a half years ago is that everything happening in the technological world and in cyberspace is increasingly central to our foreign policy. There's almost a perfect storm that's come together over the last few years, several major developments that have really brought this to the forefront of what we're doing and what we need to do. First, we have a new generation of foundational technologies that are literally changing the world all at the same time — whether it's AI, quantum, microelectronics, biotech, telecommunications. They're having a profound impact, and increasingly they're converging and feeding off of each other.

Second, we're seeing that the line between the digital and physical worlds is evaporating, erasing. We have cars, ports, hospitals that are, in effect, huge data centers. They're big vulnerabilities. At the same time, we have increasingly rare materials that are critical to technology and fragile supply chains. In each of these areas, the State Department is taking action. We have to look at everything in terms of "stacks" — the hardware, the software, the talent, and the norms, the rules, the standards by which this technology is used.

Besides setting up an entire new Bureau of Cyberspace and Digital Policy — and the bureaus are really the building blocks in our department — we've now trained more than 200 cybersecurity and digital officers, people who are genuinely expert. Every one of our embassies around the world will have at least one person who is truly fluent in tech and digital policy. My goal is to make sure that across the entire department we have basic literacy — ideally fluency — and even, eventually, mastery. All of this to make sure that, as I said, this department is fit for purpose across the entire information and digital space.
Wired notes it was Blinken's Department that discovered China's 2023 breach of Microsoft systems. And on the emerging issue of AI, Blinken cites "incredible work done by the White House to develop basic principles with the foundational companies." The voluntary commitments that they made, the State Department has worked to internationalize those commitments. We have a G7 code of conduct — the leading democratic economies in the world — all agreeing to basic principles with a focus on safety. We managed to get the very first resolution ever on artificial intelligence through the United Nations General Assembly — 192 countries also signing up to basic principles on safety and a focus on using AI to advance sustainable development goals on things like health, education, climate. We also have more than 50 countries that have signed on to basic principles on the responsible military use of AI. The goal here is not to have a world that is bifurcated in any way. It's to try to bring everyone together.

Safe Superintelligence (SSI), co-founded by OpenAI's former chief scientist Ilya Sutskever, has raised $1 billion to develop safe AI systems that surpass human capabilities. The company, valued at $5 billion, plans to use the funds to hire top talent and acquire computing power, with investors including Andreessen Horowitz, Sequoia Capital, and DST Global. Reuters reports: Sutskever, 37, is one of the most influential technologists in AI. He co-founded SSI in June with Gross, who previously led AI initiatives at Apple, and Daniel Levy, a former OpenAI researcher. Sutskever is chief scientist and Levy is principal scientist, while Gross is responsible for computing power and fundraising. Sutskever said his new venture made sense because he "identified a mountain that's a bit different from what I was working on."

SSI is currently very much focused on hiring people who will fit in with its culture. Gross said they spend hours vetting if candidates have "good character", and are looking for people with extraordinary capabilities rather than overemphasizing credentials and experience in the field. "One thing that excites us is when you find people that are interested in the work, that are not interested in the scene, in the hype," he added. SSI says it plans to partner with cloud providers and chip companies to fund its computing power needs but hasn't yet decided which firms it will work with. AI startups often work with companies such as Microsoft and Nvidia to address their infrastructure needs.

Sutskever was an early advocate of scaling, a hypothesis that AI models would improve in performance given vast amounts of computing power. The idea and its execution kicked off a wave of AI investment in chips, data centers and energy, laying the groundwork for generative AI advances like ChatGPT. Sutskever said he will approach scaling in a different way than his former employer, without sharing details. "Everyone just says scaling hypothesis. Everyone neglects to ask, what are we scaling?" he said. "Some people can work really long hours and they'll just go down the same path faster. It's not so much our style. But if you do something different, then it becomes possible for you to do something special."