Biz & IT

attackers-are-pummeling-networks-around-the-world-with-millions-of-login-attempts

Attackers are pummeling networks around the world with millions of login attempts

Biz & IT, credential compromise, Security, SSH, Talos, vpns / /

UNDER SIEGE —

Attacks coming from nearly 4,000 IP addresses take aim at VPNs, SSH and web apps.

–

Attackers are pummeling networks around the world with millions of login attempts

Matejmo | Getty Images

Cisco’s Talos security team is warning of a large-scale credential compromise campaign that’s indiscriminately assailing networks with login attempts aimed at gaining unauthorized access to VPN, SSH, and web application accounts.

The login attempts use both generic usernames and valid usernames targeted at specific organizations. Cisco included a list of more than 2,000 usernames and almost 100 passwords used in the attacks, along with nearly 4,000 IP addresses sending the login traffic. The IP addresses appear to originate from TOR exit nodes and other anonymizing tunnels and proxies. The attacks appear to be indiscriminate and opportunistic rather than aimed at a particular region or industry.

“Depending on the target environment, successful attacks of this type may lead to unauthorized network access, account lockouts, or denial-of-service conditions,” Talos researchers wrote Tuesday. “The traffic related to these attacks has increased with time and is likely to continue to rise.”

The attacks began no later than March 18.

Tuesday’s advisory comes three weeks after Cisco warned of a similar attack campaign. Cisco described that one as a password spray directed at remote access VPNs from Cisco and third-party providers connected to Cisco firewalls. This campaign appeared to be related to reconnaissance efforts, the company said.

The attacks included hundreds of thousands or millions of rejected authentication attempts. Cisco went on to say that users can intermittently receive an error message that states, “Unable to complete connection. Cisco Secure Desktop not installed on the client.” Login attempts resulting in the error fail to complete the VPN connection process. The report also reported “symptoms of hostscan token allocation failures.”

A Cisco representative said company researchers currently don’t have evidence to conclusively link the activity in both instances to the same threat actor but that there are technical overlaps in the way the attacks were carried out, as well as the infrastructure that was used.

Talos said Tuesday that services targeted in the campaign include, but aren’t limited to:

  • Cisco Secure Firewall VPN
  • Checkpoint VPN
  • Fortinet VPN
  • SonicWall VPN
  • RD Web Services
  • Mikrotik
  • Draytek
  • Ubiquiti.

Anonymization IPs appeared to belong to services, including:

  • TOR
  • VPN Gate
  • IPIDEA Proxy
  • BigMama Proxy
  • Space Proxies
  • Nexus Proxy
  • Proxy Rack.

Cisco has already added the list of IP addresses mentioned earlier to a block list for its VPN offerings. Organizations can add the addresses to block lists for any third-party VPNs they’re using. A full list of indications of compromise is here.

Cisco has also provided a list of recommendations for preventing the attacks from succeeding. The guidance includes:

  • Enabling detailed logging, ideally to a remote syslog server so that admins can recognize and correlate attacks across various network endpoints
  • Securing default remote access accounts by sinkholing them unless they use the DefaultRAGroup and DefaultWEBVPNGroup profiles
  • Blocking connection attempts from known malicious sources
  • Implement interface-level and control plane access control lists to filter out unauthorized public IP addresses and prevent them from initiating remote VPN sessions.
  • Use the shun command.

Additionally, remote access VPNs should use certificate-based authentication. Cisco lists further steps for hardening VPNs here.

Attackers are pummeling networks around the world with millions of login attempts Read More »

alleged-cryptojacking-scheme-consumed-$3.5m-of-stolen-computing-to-make-just-$1m

Alleged cryptojacking scheme consumed $3.5M of stolen computing to make just $1M

AWS, azure, Biz & IT, Cloud Computing, cryptojacking, Security / /

SHOCKING CRYPTOCURRENCY SCAM —

Indictment says man tricked cloud providers into giving him services he never paid for.

–

Alleged cryptojacking scheme consumed $3.5M of stolen computing to make just $1M

Getty Images

Federal prosecutors indicted a Nebraska man on charges he perpetrated a cryptojacking scheme that defrauded two cloud providers—one based in Seattle and the other in Redmond, Washington—out of $3.5 million.

The indictment, filed in US District Court for the Eastern District of New York and unsealed on Monday, charges Charles O. Parks III—45 of Omaha, Nebraska—with wire fraud, money laundering, and engaging in unlawful monetary transactions in connection with the scheme. Parks has yet to enter a plea and is scheduled to make an initial appearance in federal court in Omaha on Tuesday. Parks was arrested last Friday.

Prosecutors allege that Parks defrauded “two well-known providers of cloud computing services” of more than $3.5 million in computing resources to mine cryptocurrency. The indictment says the activity was in furtherance of a cryptojacking scheme, a term for crimes that generate digital coin through the acquisition of computing resources and electricity of others through fraud, hacking, or other illegal means.

Details laid out in the indictment underscore the failed economics involved in the mining of most cryptocurrencies. The $3.5 million of computing resources yielded roughly $1 million worth of cryptocurrency. In the process, massive amounts of energy were consumed.

Parks’ scheme allegedly used a variety of personal and business identities to register “numerous accounts” with the two cloud providers and in the process acquiring vast amounts of computing processing power and storage that he never paid for. Prosecutors said he tricked the providers into allotting him elevated levels of services and deferred billing accommodations and deflected the providers’ inquiries regarding questionable data usage in unpaid bills. He allegedly then used those resources to mine Ether, Litecoin, and Monero digital currencies.

The defendant then allegedly laundered the proceeds through cryptocurrency exchanges, an NFT marketplace, an online payment provider, and traditional bank accounts in an attempt to disguise the illegal scheme. Once proceeds had been converted to dollars, Parks allegedly bought a Mercedes-Benz, jewelry, first-class hotel and travel accommodations, and other luxury goods and services.

From January to August 2021, prosecutors allege, Parks created five accounts with the Seattle-based “on-demand cloud computing platform” using different names, email addresses, and corporate affiliations. He then allegedly “tricked and defrauded” employees of the platform into providing elevated levels of service, deferring billing payments, and failing to discover the activity.

During this time, Parks repeatedly requested that the provider “provide him access to powerful and expensive instances that included graphics processing units used for cryptocurrency mining and launched tens of thousands of these instances to mine cryptocurrency, employing mining software applications to facilitate the mining of tokens including ETH, LTC and XMR in various mining pools, and employing tools that allowed him to maximize cloud computing power and monitor which instances were actively mining on each mining pool,” prosecutors wrote in the indictment.

Within a day of having one account suspended for nonpayment and fraudulent activity, Parks allegedly used a new account with the provider. In all, Parks allegedly consumed more than $2.5 million of the Seattle-based provider’s services.

The prosecutors went on to allege that Parks used similar tactics to defraud the Redmond provider of more than $969,000 in cloud computing and related services.

Prosecutors didn’t say precisely how Parks was able to trick the providers into giving him elevated services, deferring unpaid payments, or failing to discover the allegedly fraudulent behavior. They also didn’t identify either of the cloud providers by name. Based on the details, however, they are almost certainly Amazon Web Services and Microsoft Azure. Representatives from both providers didn’t immediately return emails seeking confirmation.

If convicted on all charges, Parks faces as much as 30 years in prison.

Alleged cryptojacking scheme consumed $3.5M of stolen computing to make just $1M Read More »

framework’s-software-and-firmware-have-been-a-mess,-but-it’s-working-on-them

Framework’s software and firmware have been a mess, but it’s working on them

AMD, BIOS, Biz & IT, Features, framework laptop, framework laptop 13, framework laptop 16, intel, Security, Tech / /
The Framework Laptop 13.

Enlarge / The Framework Laptop 13.

Andrew Cunningham

Since Framework showed off its first prototypes in February 2021, we’ve generally been fans of the company’s modular, repairable, upgradeable laptops.

Not that the company’s hardware releases to date have been perfect—each Framework Laptop 13 model has had quirks and flaws that range from minor to quite significant, and the Laptop 16’s upsides struggle to balance its downsides. But the hardware mostly does a good job of functioning as a regular laptop while being much more tinkerer-friendly than your typical MacBook, XPS, or ThinkPad.

But even as it builds new upgrades for its systems, expands sales of refurbished and B-stock hardware as budget options, and promotes the re-use of its products via external enclosures, Framework has struggled with the other side of computing longevity and sustainability: providing up-to-date software.

Driver bundles remain un-updated for years after their initial release. BIOS updates go through long and confusing beta processes, keeping users from getting feature improvements, bug fixes, and security updates. In its community support forums, Framework employees, including founder and CEO Nirav Patel, have acknowledged these issues and promised fixes but have remained inconsistent and vague about actual timelines.

But according to Patel, the company is working on fixing these issues, and it has taken some steps to address them. We spoke to him about the causes of and the solutions to these issues, and the company’s approach to the software side of its efforts to promote repairability and upgradeability.

Promises made

Here’s a case in point: the 12th-generation Intel version of the Framework Laptop 13, which prompted me to start monitoring Framework’s software and firmware updates in the first place.

In November 2022, Patel announced that this model, then the latest version, was getting a nice, free-of-charge spec bump. All four of the laptop’s recessed USB-C ports would now become full-speed Thunderbolt ports. This wasn’t a dramatic functional change, especially for people who were mostly using those ports for basic Framework expansion modules like USB-A or HDMI, but the upgrade opened the door to high-speed external accessories, and all it would need was a BIOS update.

The recessed USB-C ports in the 12th-gen Intel version of the Framework Laptop 13 can be upgraded to fully certified Thunderbolt ports, but only if you're willing to install one in a long series of still-in-testing beta BIOSes.

Enlarge / The recessed USB-C ports in the 12th-gen Intel version of the Framework Laptop 13 can be upgraded to fully certified Thunderbolt ports, but only if you’re willing to install one in a long series of still-in-testing beta BIOSes.

Andrew Cunningham

A final version of this BIOS update finally showed up this week, nearly a year and a half later. Up until last week, Framework’s support page for that 12th-gen Intel laptop still said that there was “no new BIOS available” for a laptop that began shipping in the summer of 2022. This factory-installed BIOS, version 3.04, also didn’t include fixes for the LogoFAIL UEFI security vulnerability or any other firmware-based security patches that have cropped up in the last year and a half.

And it’s not just that the updates don’t come out in a timely way; the company has been bad about estimating when they might come out. That old12th-gen Framework BIOS also didn’t support the 61 WHr battery that the company released in early 2023 alongside the 13th-gen Intel refresh. Framework originally told me that BIOS update would be out in May of 2023. A battery-supporting update for the 11th-gen Intel version was also promised in May 2023; it came out this past January.

Framework has been trying, but it keeps running into issues. A beta 3.06 BIOS update with the promised improvements for the 12th-gen Intel Framework Laptop was posted back in December of 2022, but a final version was never released. The newer 3.08 BIOS beta entered testing in January 2024 but still gave users some problems. Users would go for weeks or months without any communication from anyone at Framework.

The result is multiple long forum threads of frustrated users asking for updates, interspersed with not-untrue but unsatisfying responses from Framework employees (some version of “we’re a small company” is one of the most common).

Framework’s software and firmware have been a mess, but it’s working on them Read More »

change-healthcare-faces-another-ransomware-threat—and-it-looks-credible

Change Healthcare faces another ransomware threat—and it looks credible

Biz & IT, Security / /
Medical Data Breach text write on keyboard isolated on laptop background

For months, Change Healthcare has faced an immensely messy ransomware debacle that has left hundreds of pharmacies and medical practices across the United States unable to process claims. Now, thanks to an apparent dispute within the ransomware criminal ecosystem, it may have just become far messier still.

In March, the ransomware group AlphV, which had claimed credit for encrypting Change Healthcare’s network and threatened to leak reams of the company’s sensitive health care data, received a $22 million payment—evidence, publicly captured on bitcoin’s blockchain, that Change Healthcare had very likely caved to its tormentors’ ransom demand, though the company has yet to confirm that it paid. But in a new definition of a worst-case ransomware, a different ransomware group claims to be holding Change Healthcare’s stolen data and is demanding a payment of their own.

Since Monday, RansomHub, a relatively new ransomware group, has posted to its dark-web site that it has 4 terabytes of Change Healthcare’s stolen data, which it threatened to sell to the “highest bidder” if Change Healthcare didn’t pay an unspecified ransom. RansomHub tells WIRED it is not affiliated with AlphV and “can’t say” how much it’s demanding as a ransom payment.

RansomHub initially declined to publish or provide WIRED any sample data from that stolen trove to prove its claim. But on Friday, a representative for the group sent WIRED several screenshots of what appeared to be patient records and a data-sharing contract for United Healthcare, which owns Change Healthcare, and Emdeon, which acquired Change Healthcare in 2014 and later took its name.

While WIRED could not fully confirm RansomHub’s claims, the samples suggest that this second extortion attempt against Change Healthcare may be more than an empty threat. “For anyone doubting that we have the data, and to anyone speculating the criticality and the sensitivity of the data, the images should be enough to show the magnitude and importance of the situation and clear the unrealistic and childish theories,” the RansomHub contact tells WIRED in an email.

Change Healthcare didn’t immediately respond to WIRED’s request for comment on RansomHub’s extortion demand.

Brett Callow, a ransomware analyst with security firm Emsisoft, says he believes AlphV did not originally publish any data from the incident, and the origin of RansomHub’s data is unclear. “I obviously don’t know whether the data is real—it could have been pulled from elsewhere—but nor do I see anything that indicates it may not be authentic,” he says of the data shared by RansomHub.

Jon DiMaggio, chief security strategist at threat intelligence firm Analyst1, says he believes RansomHub is “telling the truth and does have Change HealthCare’s data,” after reviewing the information sent to WIRED. While RansomHub is a new ransomware threat actor, DiMaggio says, they are quickly “gaining momentum.”

If RansomHub’s claims are real, it will mean that Change Healthcare’s already catastrophic ransomware ordeal has become a kind of cautionary tale about the dangers of trusting ransomware groups to follow through on their promises, even after a ransom is paid. In March, someone who goes by the name “notchy” posted to a Russian cybercriminal forum that AlphV had pocketed that $22 million payment and disappeared without sharing a commission with the “affiliate” hackers who typically partner with ransomware groups and often penetrate victims’ networks on their behalf.

Change Healthcare faces another ransomware threat—and it looks credible Read More »

“highly-capable”-hackers-root-corporate-networks-by-exploiting-firewall-0-day

“Highly capable” hackers root corporate networks by exploiting firewall 0-day

Biz & IT, exploits, Palo Alto Networks, pan-os, Security, vulnerabilities, zerodays / /
The word ZERO-DAY is hidden amidst a screen filled with ones and zeroes.

Highly capable hackers are rooting multiple corporate networks by exploiting a maximum-severity zero-day vulnerability in a firewall product from Palo Alto Networks, researchers said Friday.

The vulnerability, which has been under active exploitation for at least two weeks now, allows the hackers with no authentication to execute malicious code with root privileges, the highest possible level of system access, researchers said. The extent of the compromise, along with the ease of exploitation, has earned the CVE-2024-3400 vulnerability the maximum severity rating of 10.0. The ongoing attacks are the latest in a rash of attacks aimed at firewalls, VPNs, and file-transfer appliances, which are popular targets because of their wealth of vulnerabilities and direct pipeline into the most sensitive parts of a network.

“Highly capable” UTA0218 likely to be joined by others

The zero-day is present in PAN-OS 10.2, PAN-OS 11.0, and/or PAN-OS 11.1 firewalls when they are configured to use both the GlobalProtect gateway and device telemetry. Palo Alto Networks has yet to patch the vulnerability but is urging affected customers to follow the workaround and mitigation guidance provided here. The advice includes enabling Threat ID 95187 for those with subscriptions to the company’s Threat Prevention service and ensuring vulnerability protection has been applied to their GlobalProtect interface. When that’s not possible, customers should temporarily disable telemetry until a patch is available.

Volexity, the security firm that discovered the zero-day attacks, said that it’s currently unable to tie the attackers to any previously known groups. However, based on the resources required and the organizations targeted, they are “highly capable” and likely backed by a nation-state. So far, only a single threat group—which Volexity tracks as UTA0218—is known to be leveraging the vulnerability in limited attacks. The company warned that as new groups learn of the vulnerability, CVE-2024-3400, is likely to come under mass exploitation, just as recent zero-days affecting products from the likes of Ivanti, Atlassian, Citrix, and Progress have in recent months.

“As with previous public disclosures of vulnerabilities in these kinds of devices, Volexity assesses that it is likely a spike in exploitation will be observed over the next few days by UTA0218 and potentially other threat actors who may develop exploits for this vulnerability,” company researchers wrote Friday. “This spike in activity will be driven by the urgency of this window of access closing due to mitigations and patches being deployed. It is therefore imperative that organizations act quickly to deploy recommended mitigations and perform compromise reviews of their devices to check whether further internal investigation of their networks is required.”

The earliest attacks Volexity has seen took place on March 26 in what company researchers suspect was UTA0218 testing the vulnerability by placing zero-byte files on firewall devices to validate exploitability. On April 7, the researchers observed the group trying unsuccessfully to install a backdoor on a customer’s firewall. Three days later, the group’s attacks were successfully deploying malicious payloads. Since then, the threat group has deployed custom, never-before-seen post-exploitation malware. The backdoor, which is written in the Python language, allows the attackers to use specially crafted network requests to execute additional commands on hacked devices.

“Highly capable” hackers root corporate networks by exploiting firewall 0-day Read More »

words-are-flowing-out-like-endless-rain:-recapping-a-busy-week-of-llm-news

Words are flowing out like endless rain: Recapping a busy week of LLM news

AI, Anthropic, Biz & IT, Chatbot Arena, chatgpt, chatgtp, Claude 3, Claude 3 Opus, Claude 3 Sonnet, Gemini, Gemini 1.5, Google Gemini, GPT-4, GPT-4-turbo, large language models, machine learning, Meta, Mistral, Mixtral, Mixtral 8x22B, openai, RECAP, Simon Willison, whirlwind / /

many things frequently —

Gemini 1.5 Pro launch, new version of GPT-4 Turbo, new Mistral model, and more.

–

An image of a boy amazed by flying letters.

Enlarge / An image of a boy amazed by flying letters.

Some weeks in AI news are eerily quiet, but during others, getting a grip on the week’s events feels like trying to hold back the tide. This week has seen three notable large language model (LLM) releases: Google Gemini Pro 1.5 hit general availability with a free tier, OpenAI shipped a new version of GPT-4 Turbo, and Mistral released a new openly licensed LLM, Mixtral 8x22B. All three of those launches happened within 24 hours starting on Tuesday.

With the help of software engineer and independent AI researcher Simon Willison (who also wrote about this week’s hectic LLM launches on his own blog), we’ll briefly cover each of the three major events in roughly chronological order, then dig into some additional AI happenings this week.

Gemini Pro 1.5 general release

On Tuesday morning Pacific time, Google announced that its Gemini 1.5 Pro model (which we first covered in February) is now available in 180-plus countries, excluding Europe, via the Gemini API in a public preview. This is Google’s most powerful public LLM so far, and it’s available in a free tier that permits up to 50 requests a day.

It supports up to 1 million tokens of input context. As Willison notes in his blog, Gemini 1.5 Pro’s API price at $7/million input tokens and $21/million output tokens costs a little less than GPT-4 Turbo (priced at $10/million in and $30/million out) and more than Claude 3 Sonnet (Anthropic’s mid-tier LLM, priced at $3/million in and $15/million out).

Notably, Gemini 1.5 Pro includes native audio (speech) input processing that allows users to upload audio or video prompts, a new File API for handling files, the ability to add custom system instructions (system prompts) for guiding model responses, and a JSON mode for structured data extraction.

“Majorly Improved” GPT-4 Turbo launch

A GPT-4 Turbo performance chart provided by OpenAI.

Enlarge / A GPT-4 Turbo performance chart provided by OpenAI.

Just a bit later than Google’s 1.5 Pro launch on Tuesday, OpenAI announced that it was rolling out a “majorly improved” version of GPT-4 Turbo (a model family originally launched in November) called “gpt-4-turbo-2024-04-09.” It integrates multimodal GPT-4 Vision processing (recognizing the contents of images) directly into the model, and it initially launched through API access only.

Then on Thursday, OpenAI announced that the new GPT-4 Turbo model had just become available for paid ChatGPT users. OpenAI said that the new model improves “capabilities in writing, math, logical reasoning, and coding” and shared a chart that is not particularly useful in judging capabilities (that they later updated). The company also provided an example of an alleged improvement, saying that when writing with ChatGPT, the AI assistant will use “more direct, less verbose, and use more conversational language.”

The vague nature of OpenAI’s GPT-4 Turbo announcements attracted some confusion and criticism online. On X, Willison wrote, “Who will be the first LLM provider to publish genuinely useful release notes?” In some ways, this is a case of “AI vibes” again, as we discussed in our lament about the poor state of LLM benchmarks during the debut of Claude 3. “I’ve not actually spotted any definite differences in quality [related to GPT-4 Turbo],” Willison told us directly in an interview.

The update also expanded GPT-4’s knowledge cutoff to April 2024, although some people are reporting it achieves this through stealth web searches in the background, and others on social media have reported issues with date-related confabulations.

Mistral’s mysterious Mixtral 8x22B release

An illustration of a robot holding a French flag, figuratively reflecting the rise of AI in France due to Mistral. It's hard to draw a picture of an LLM, so a robot will have to do.

Enlarge / An illustration of a robot holding a French flag, figuratively reflecting the rise of AI in France due to Mistral. It’s hard to draw a picture of an LLM, so a robot will have to do.

Not to be outdone, on Tuesday night, French AI company Mistral launched its latest openly licensed model, Mixtral 8x22B, by tweeting a torrent link devoid of any documentation or commentary, much like it has done with previous releases.

The new mixture-of-experts (MoE) release weighs in with a larger parameter count than its previously most-capable open model, Mixtral 8x7B, which we covered in December. It’s rumored to potentially be as capable as GPT-4 (In what way, you ask? Vibes). But that has yet to be seen.

“The evals are still rolling in, but the biggest open question right now is how well Mixtral 8x22B shapes up,” Willison told Ars. “If it’s in the same quality class as GPT-4 and Claude 3 Opus, then we will finally have an openly licensed model that’s not significantly behind the best proprietary ones.”

This release has Willison most excited, saying, “If that thing really is GPT-4 class, it’s wild, because you can run that on a (very expensive) laptop. I think you need 128GB of MacBook RAM for it, twice what I have.”

The new Mixtral is not listed on Chatbot Arena yet, Willison noted, because Mistral has not released a fine-tuned model for chatting yet. It’s still a raw, predict-the-next token LLM. “There’s at least one community instruction tuned version floating around now though,” says Willison.

Chatbot Arena Leaderboard shake-ups

A Chatbot Arena Leaderboard screenshot taken on April 12, 2024.

Enlarge / A Chatbot Arena Leaderboard screenshot taken on April 12, 2024.

Benj Edwards

This week’s LLM news isn’t limited to just the big names in the field. There have also been rumblings on social media about the rising performance of open source models like Cohere’s Command R+, which reached position 6 on the LMSYS Chatbot Arena Leaderboard—the highest-ever ranking for an open-weights model.

And for even more Chatbot Arena action, apparently the new version of GPT-4 Turbo is proving competitive with Claude 3 Opus. The two are still in a statistical tie, but GPT-4 Turbo recently pulled ahead numerically. (In March, we reported when Claude 3 first numerically pulled ahead of GPT-4 Turbo, which was then the first time another AI model had surpassed a GPT-4 family model member on the leaderboard.)

Regarding this fierce competition among LLMs—of which most of the muggle world is unaware and will likely never be—Willison told Ars, “The past two months have been a whirlwind—we finally have not just one but several models that are competitive with GPT-4.” We’ll see if OpenAI’s rumored release of GPT-5 later this year will restore the company’s technological lead, we note, which once seemed insurmountable. But for now, Willison says, “OpenAI are no longer the undisputed leaders in LLMs.”

Words are flowing out like endless rain: Recapping a busy week of LLM news Read More »

intel’s-“gaudi-3”-ai-accelerator-chip-may-give-nvidia’s-h100-a-run-for-its-money

Intel’s “Gaudi 3” AI accelerator chip may give Nvidia’s H100 a run for its money

AI, AI chips, B200, Biz & IT, Blackwell, chatgpt, chatgtp, Gaudi 3, H100, H200, intel, Intel Gaudi, machine learning, NVIDIA, Nvidia Blackwell, openai / /

Adventures in Matrix Multiplication —

Intel claims 50% more speed when running AI language models vs. the market leader.

–

An Intel handout photo of the Gaudi 3 AI accelerator.

Enlarge / An Intel handout photo of the Gaudi 3 AI accelerator.

On Tuesday, Intel revealed a new AI accelerator chip called Gaudi 3 at its Vision 2024 event in Phoenix. With strong claimed performance while running large language models (like those that power ChatGPT), the company has positioned Gaudi 3 as an alternative to Nvidia’s H100, a popular data center GPU that has been subject to shortages, though apparently that is easing somewhat.

Compared to Nvidia’s H100 chip, Intel projects a 50 percent faster training time on Gaudi 3 for both OpenAI’s GPT-3 175B LLM and the 7-billion parameter version of Meta’s Llama 2. In terms of inference (running the trained model to get outputs), Intel claims that its new AI chip delivers 50 percent faster performance than H100 for Llama 2 and Falcon 180B, which are both relatively popular open-weights models.

Intel is targeting the H100 because of its high market share, but the chip isn’t Nvidia’s most powerful AI accelerator chip in the pipeline. Announcements of the H200 and the Blackwell B200 have since surpassed the H100 on paper, but neither of those chips is out yet (the H200 is expected in the second quarter of 2024—basically any day now).

Meanwhile, the aforementioned H100 supply issues have been a major headache for tech companies and AI researchers who have to fight for access to any chips that can train AI models. This has led several tech companies like Microsoft, Meta, and OpenAI (rumor has it) to seek their own AI-accelerator chip designs, although that custom silicon is typically manufactured by either Intel or TSMC. Google has its own line of tensor processing units (TPUs) that it has been using internally since 2015.

Given those issues, Intel’s Gaudi 3 may be a potentially attractive alternative to the H100 if Intel can hit an ideal price (which Intel has not provided, but an H100 reportedly costs around $30,000–$40,000) and maintain adequate production. AMD also manufactures a competitive range of AI chips, such as the AMD Instinct MI300 Series, that sell for around $10,000–$15,000.

Gaudi 3 performance

An Intel handout featuring specifications of the Gaudi 3 AI accelerator.

Enlarge / An Intel handout featuring specifications of the Gaudi 3 AI accelerator.

Intel says the new chip builds upon the architecture of its predecessor, Gaudi 2, by featuring two identical silicon dies connected by a high-bandwidth connection. Each die contains a central cache memory of 48 megabytes, surrounded by four matrix multiplication engines and 32 programmable tensor processor cores, bringing the total cores to 64.

The chipmaking giant claims that Gaudi 3 delivers double the AI compute performance of Gaudi 2 using 8-bit floating-point infrastructure, which has become crucial for training transformer models. The chip also offers a fourfold boost for computations using the BFloat 16-number format. Gaudi 3 also features 128GB of the less expensive HBMe2 memory capacity (which may contribute to price competitiveness) and features 3.7TB of memory bandwidth.

Since data centers are well-known to be power hungry, Intel emphasizes the power efficiency of Gaudi 3, claiming 40 percent greater inference power-efficiency across Llama 7B and 70B parameters, and Falcon 180B parameter models compared to Nvidia’s H100. Eitan Medina, chief operating officer of Intel’s Habana Labs, attributes this advantage to Gaudi’s large-matrix math engines, which he claims require significantly less memory bandwidth compared to other architectures.

Gaudi vs. Blackwell

An Intel handout photo of the Gaudi 3 AI accelerator.

Enlarge / An Intel handout photo of the Gaudi 3 AI accelerator.

Last month, we covered the splashy launch of Nvidia’s Blackwell architecture, including the B200 GPU, which Nvidia claims will be the world’s most powerful AI chip. It seems natural, then, to compare what we know about Nvidia’s highest-performing AI chip to the best of what Intel can currently produce.

For starters, Gaudi 3 is being manufactured using TSMC’s N5 process technology, according to IEEE Spectrum, narrowing the gap between Intel and Nvidia in terms of semiconductor fabrication technology. The upcoming Nvidia Blackwell chip will use a custom N4P process, which reportedly offers modest performance and efficiency improvements over N5.

Gaudi 3’s use of HBM2e memory (as we mentioned above) is notable compared to the more expensive HBM3 or HBM3e used in competing chips, offering a balance of performance and cost-efficiency. This choice seems to emphasize Intel’s strategy to compete not only on performance but also on price.

As far as raw performance comparisons between Gaudi 3 and the B200, that can’t be known until the chips have been released and benchmarked by a third party.

As the race to power the tech industry’s thirst for AI computation heats up, IEEE Spectrum notes that the next generation of Intel’s Gaudi chip, code-named Falcon Shores, remains a point of interest. It also remains to be seen whether Intel will continue to rely on TSMC’s technology or leverage its own foundry business and upcoming nanosheet transistor technology to gain a competitive edge in the AI accelerator market.

Intel’s “Gaudi 3” AI accelerator chip may give Nvidia’s H100 a run for its money Read More »

at&t:-data-breach-affects-73-million-or-51-million-customers-no,-we-won’t-explain.

AT&T: Data breach affects 73 million or 51 million customers. No, we won’t explain.

AT&T, Biz & IT, Data breaches, Identity theft, Security / /

“SECURITY IS IMPORTANT TO US” —

When the data was published in 2021, the company said it didn’t belong to its customers.

–

AT&T: Data breach affects 73 million or 51 million customers. No, we won’t explain.

Getty Images

AT&T is notifying millions of current or former customers that their account data has been compromised and published last month on the dark web. Just how many millions, the company isn’t saying.

In a mandatory filing with the Maine Attorney General’s office, the telecommunications company said 51.2 million account holders were affected. On its corporate website, AT&T put the number at 73 million. In either event, compromised data included one or more of the following: full names, email addresses, mailing addresses, phone numbers, social security numbers, dates of birth, AT&T account numbers, and AT&T passcodes. Personal financial information and call history didn’t appear to be included, AT&T said, and data appeared to be from June 2019 or earlier.

The disclosure on the AT&T site said the 73 million affected customers comprised 7.6 million current customers and 65.4 million former customers. The notification said AT&T has reset the account PINs of all current customers and is notifying current and former customers by mail. AT&T representatives haven’t explained why the letter filed with the Maine AG lists 51.2 million affected and the disclosure on its site lists 73 million.

According to a March 30 article published by TechCrunch, a security researcher said the passcodes were stored in an encrypted format that could easily be decrypted. Bleeping Computer reported in 2021 that more than 70 million records containing AT&T customer data was put up for sale that year for $1 million. AT&T, at the time, told the news site that the amassed data didn’t belong to its customers and that the company’s systems had not been breached.

Last month, after the same data reappeared online, Bleeping Computer and TechCrunch confirmed that the data belonged to AT&T customers, and the company finally acknowledged the connection. AT&T has yet to say how the information was breached or why it took more than two years from the original date of publication to confirm that it belonged to its customers.

Given the length of time the data has been available, the damage that’s likely to result from the most recent publication is likely to be minimal. That said, anyone who is or was an AT&T customer should be on the lookout for scams that attempt to capitalize on the leaked data. AT&T is offering one year of free identity theft protection.

AT&T: Data breach affects 73 million or 51 million customers. No, we won’t explain. Read More »

new-ai-music-generator-udio-synthesizes-realistic-music-on-demand

New AI music generator Udio synthesizes realistic music on demand

AI, audio synthesis, Biz & IT, chatgpt, chatgtp, machine learning, music, music synthesis, openai, Suno, Udio / /

Battle of the AI bands —

But it still needs trial and error to generate high-quality results.

–

A screenshot of AI-generated songs listed on Udio on April 10, 2024.

Enlarge / A screenshot of AI-generated songs listed on Udio on April 10, 2024.

Benj Edwards

Between 2002 and 2005, I ran a music website where visitors could submit song titles that I would write and record a silly song around. In the liner notes for my first CD release in 2003, I wrote about a day when computers would potentially put me out of business, churning out music automatically at a pace I could not match. While I don’t actively post music on that site anymore, that day is almost here.

On Wednesday, a group of ex-DeepMind employees launched Udio, a new AI music synthesis service that can create novel high-fidelity musical audio from written prompts, including user-provided lyrics. It’s similar to Suno, which we covered on Monday. With some key human input, Udio can create facsimiles of human-produced music in genres like country, barbershop quartet, German pop, classical, hard rock, hip hop, show tunes, and more. It’s currently free to use during a beta period.

Udio is also freaking out some musicians on Reddit. As we mentioned in our Suno piece, Udio is exactly the kind of AI-powered music generation service that over 200 musical artists were afraid of when they signed an open protest letter last week.

But as impressive as the Udio songs first seem from a technical AI-generation standpoint (not necessarily judging by musical merit), its generation capability isn’t perfect. We experimented with its creation tool and the results felt less impressive than those created by Suno. The high-quality musical samples showcased on Udio’s site likely resulted from a lot of creative human input (such as human-written lyrics) and cherry-picking the best compositional parts of songs out of many generations. In fact, Udio lays out a five-step workflow to build a 1.5-minute-long song in a FAQ.

For example, we created an Ars Technica “Moonshark” song on Udio using the same prompt as one we used previously with Suno. In its raw form, the results sound half-baked and almost nightmarish (here is the Suno version for comparison). It’s also a lot shorter by default at 32 seconds compared to Suno’s 1-minute and 32-second output. But Udio allows songs to be extended, or you can try generating a poor result again with different prompts for different results.

After registering a Udio account, anyone can create a track by entering a text prompt that can include lyrics, a story direction, and musical genre tags. Udio then tackles the task in two stages. First, it utilizes a large language model (LLM) similar to ChatGPT to generate lyrics (if necessary) based on the provided prompt. Next, it synthesizes music using a method that Udio does not disclose, but it’s likely a diffusion model, similar to Stability AI’s Stable Audio.

From the given prompt, Udio’s AI model generates two distinct song snippets for you to choose from. You can then publish the song for the Udio community, download the audio or video file to share on other platforms, or directly share it on social media. Other Udio users can also remix or build on existing songs. Udio’s terms of service say that the company claims no rights over the musical generations and that they can be used for commercial purposes.

Although the Udio team has not revealed the specific details of its model or training data (which is likely filled with copyrighted material), it told Tom’s Guide that the system has built-in measures to identify and block tracks that too closely resemble the work of specific artists, ensuring that the generated music remains original.

And that brings us back to humans, some of whom are not taking the onset of AI-generated music very well. “I gotta be honest, this is depressing as hell,” wrote one Reddit commenter in a thread about Udio. “I’m still broadly optimistic that music will be fine in the long run somehow. But like, why do this? Why automate art?”

We’ll hazard an answer by saying that replicating art is a key target for AI research because the results can be inaccurate and imprecise and still seem notable or gee-whiz amazing, which is a key characteristic of generative AI. It’s flashy and impressive-looking while allowing for a general lack of quantitative rigor. We’ve already seen AI come for still images, video, and text with varied results regarding representative accuracy. Fully composed musical recordings seem to be next on the list of AI hills to (approximately) conquer, and the competition is heating up.

New AI music generator Udio synthesizes realistic music on demand Read More »

thousands-of-lg-tvs-are-vulnerable-to-takeover—here’s-how-to-ensure-yours-isn’t-one

Thousands of LG TVs are vulnerable to takeover—here’s how to ensure yours isn’t one

Biz & IT, hacking, Internet of things, lg, Security, TVs / /
Thousands of LG TVs are vulnerable to takeover—here’s how to ensure yours isn’t one

Getty Images

As many as 91,000 LG TVs face the risk of being commandeered unless they receive a just-released security update patching four critical vulnerabilities discovered late last year.

The vulnerabilities are found in four LG TV models that collectively comprise slightly more than 88,000 units around the world, according to results returned by the Shodan search engine for Internet-connected devices. The vast majority of those units are located in South Korea, followed by Hong Kong, the US, Sweden, and Finland. The models are:

  • LG43UM7000PLA running webOS 4.9.7 – 5.30.40
  • OLED55CXPUA running webOS 5.5.0 – 04.50.51
  • OLED48C1PUB running webOS 6.3.3-442 (kisscurl-kinglake) – 03.36.50
  • OLED55A23LA running webOS 7.3.1-43 (mullet-mebin) – 03.33.85

Starting Wednesday, updates are available through these devices’ settings menu.

Got root?

According to Bitdefender—the security firm that discovered the vulnerabilities—malicious hackers can exploit them to gain root access to the devices and inject commands that run at the OS level. The vulnerabilities, which affect internal services that allow users to control their sets using their phones, make it possible for attackers to bypass authentication measures designed to ensure only authorized devices can make use of the capabilities.

“These vulnerabilities let us gain root access on the TV after bypassing the authorization mechanism,” Bitdefender researchers wrote Tuesday. “Although the vulnerable service is intended for LAN access only, Shodan, the search engine for Internet-connected devices, identified over 91,000 devices that expose this service to the Internet.”

The key vulnerability making these threats possible resides in a service that allows TVs to be controlled using LG’s ThinkQ smartphone app when it’s connected to the same local network. The service is designed to require the user to enter a PIN code to prove authorization, but an error allows someone to skip this verification step and become a privileged user. This vulnerability is tracked as CVE-2023-6317.

Once attackers have gained this level of control, they can go on to exploit three other vulnerabilities, specifically:

  • CVE-2023-6318, which allows the attackers to elevate their access to root
  • CVE-2023-6319, which allows for the injection of OS commands by manipulating a library for showing music lyrics
  • CVE-2023-6320, which lets an attacker inject authenticated commands by manipulating the com.webos.service.connectionmanager/tv/setVlanStaticAddress application interface.

Thousands of LG TVs are vulnerable to takeover—here’s how to ensure yours isn’t one Read More »

elon-musk:-ai-will-be-smarter-than-any-human-around-the-end-of-next-year

Elon Musk: AI will be smarter than any human around the end of next year

AI, Biz & IT, elon musk, Grady Booch, machine learning, Social Media, superintelligence, Twitter, X / /

smarter than the average bear —

While Musk says superintelligence is coming soon, one critic says prediction is “batsh*t crazy.”

–

Elon Musk, owner of Tesla and the X (formerly Twitter) platform, attends a symposium on fighting antisemitism titled 'Never Again : Lip Service or Deep Conversation' in Krakow, Poland on January 22nd, 2024. Musk, who was invited to Poland by the European Jewish Association (EJA) has visited the Auschwitz-Birkenau concentration camp earlier that day, ahead of International Holocaust Remembrance Day. (Photo by Beata Zawrzel/NurPhoto)

Enlarge / Elon Musk, owner of Tesla and the X (formerly Twitter) platform on January 22, 2024.

On Monday, Tesla CEO Elon Musk predicted the imminent rise in AI superintelligence during a live interview streamed on the social media platform X. “My guess is we’ll have AI smarter than any one human probably around the end of next year,” Musk said in his conversation with hedge fund manager Nicolai Tangen.

Just prior to that, Tangen had asked Musk, “What’s your take on where we are in the AI race just now?” Musk told Tangen that AI “is the fastest advancing technology I’ve seen of any kind, and I’ve seen a lot of technology.” He described computers dedicated to AI increasing in capability by “a factor of 10 every year, if not every six to nine months.”

Musk made the prediction with an asterisk, saying that shortages of AI chips and high AI power demands could limit AI’s capability until those issues are resolved. “Last year, it was chip-constrained,” Musk told Tangen. “People could not get enough Nvidia chips. This year, it’s transitioning to a voltage transformer supply. In a year or two, it’s just electricity supply.”

But not everyone is convinced that Musk’s crystal ball is free of cracks. Grady Booch, a frequent critic of AI hype on social media who is perhaps best known for his work in software architecture, told Ars in an interview, “Keep in mind that Mr. Musk has a profoundly bad record at predicting anything associated with AI; back in 2016, he promised his cars would ship with FSD safety level 5, and here we are, closing on an a decade later, still waiting.”

Creating artificial intelligence at least as smart as a human (frequently called “AGI” for artificial general intelligence) is often seen as inevitable among AI proponents, but there’s no broad consensus on exactly when that milestone will be reached—or on the exact definition of AGI, for that matter.

“If you define AGI as smarter than the smartest human, I think it’s probably next year, within two years,” Musk added in the interview with Tangen while discussing AGI timelines.

Even with uncertainties about AGI, that hasn’t kept companies from trying. ChatGPT creator OpenAI, which launched with Musk as a co-founder in 2015, lists developing AGI as its main goal. Musk has not been directly associated with OpenAI for years (unless you count a recent lawsuit against the company), but last year, he took aim at the business of large language models by forming a new company called xAI. Its main product, Grok, functions similarly to ChatGPT and is integrated into the X social media platform.

Booch gives credit to Musk’s business successes but casts doubt on his forecasting ability. “Albeit a brilliant if not rapacious businessman, Mr. Musk vastly overestimates both the history as well as the present of AI while simultaneously diminishing the exquisite uniqueness of human intelligence,” says Booch. “So in short, his prediction is—to put it in scientific terms—batshit crazy.”

So when will we get AI that’s smarter than a human? Booch says there’s no real way to know at the moment. “I reject the framing of any question that asks when AI will surpass humans in intelligence because it is a question filled with ambiguous terms and considerable emotional and historic baggage,” he says. “We are a long, long way from understanding the design that would lead us there.”

We also asked Hugging Face AI researcher Dr. Margaret Mitchell to weigh in on Musk’s prediction. “Intelligence … is not a single value where you can make these direct comparisons and have them mean something,” she told us in an interview. “There will likely never be agreement on comparisons between human and machine intelligence.”

But even with that uncertainty, she feels there is one aspect of AI she can more reliably predict: “I do agree that neural network models will reach a point where men in positions of power and influence, particularly ones with investments in AI, will declare that AI is smarter than humans. By end of next year, sure. That doesn’t sound far off base to me.”

Elon Musk: AI will be smarter than any human around the end of next year Read More »

mit-license-text-becomes-viral-“sad-girl”-piano-ballad-generated-by-ai

MIT License text becomes viral “sad girl” piano ballad generated by AI

AI, AI-generated music, barbarians, Biz & IT, chatgpt, chatgtp, machine learning, moonshark, music synthesis, openai, Riley Goodside, sound synthesis, Suno, Suno.ai / /

WARRANTIES OF MERCHANTABILITY —

“Permission is hereby granted” comes from Suno AI engine that creates new songs on demand.

–

Illustration of a robot singing.

We’ve come a long way since primitive AI music generators in 2022. Today, AI tools like Suno.ai allow any series of words to become song lyrics, including inside jokes (as you’ll see below). On Wednesday, prompt engineer Riley Goodside tweeted an AI-generated song created with the prompt “sad girl with piano performs the text of the MIT License,” and it began to circulate widely in the AI community online.

The MIT License is a famous permissive software license created in the late 1980s, frequently used in open source projects. “My favorite part of this is ~1: 25 it nails ‘WARRANTIES OF MERCHANTABILITY’ with a beautiful Imogen Heap-style glissando then immediately pronounces ‘FITNESS’ as ‘fistiff,'” Goodside wrote on X.

Suno (which means “listen” in Hindi) was formed in 2023 in Cambridge, Massachusetts. It’s the brainchild of Michael Shulman, Georg Kucsko, Martin Camacho, and Keenan Freyberg, who formerly worked at companies like Meta and TikTok. Suno has already attracted big-name partners, such as Microsoft, which announced the integration of an earlier version of the Suno engine into Bing Chat last December. Today, Suno is on v3 of its model, which can create temporally coherent two-minute songs in many different genres.

The company did not reply to our request for an interview by press time. In March, Brian Hiatt of Rolling Stone wrote a profile about Suno that describes the service as a collaboration between OpenAI’s ChatGPT (for lyric writing) and Suno’s music generation model, which some experts think has likely been trained on recordings of copyrighted music without license or artist permission.

It’s exactly this kind of service that upset over 200 musical artists enough last week that they signed an Artist Rights Alliance open letter asking tech companies to stop using AI tools to generate music that could replace human artists.

Considering the unknown provenance of the training data, ownership of the generated songs seems like a complicated question. Suno’s FAQ says that music generated using its free tier remains owned by Suno and can only be used for non-commercial purposes. Paying subscribers reportedly own generated songs “while subscribed to Pro or Premier,” subject to Suno’s terms of service. However, the US Copyright Office took a stance last year that purely AI-generated visual art cannot be copyrighted, and while that standard has not yet been resolved for AI-generated music, it might eventually become official legal policy as well.

The Moonshark song

A screenshot of the Suno.ai website showing lyrics of an AI-generated

Enlarge / A screenshot of the Suno.ai website showing lyrics of an AI-generated “Moonshark” song.

Benj Edwards

While using the service, Suno appears to have no trouble creating unique lyrics based on your prompt (unless you supply your own) and sets those words to stylized genres of music it generates based on its training dataset. It dynamically generates vocals as well, although they include audible aberrations. Suno’s output is not indistinguishable from high-fidelity human-created music yet, but given the pace of progress we’ve seen, that bridge could be crossed within the next year.

To get a sense of what Suno can do, we created an account on the site and prompted the AI engine to create songs about our mascot, Moonshark, and about barbarians with CRTs, two inside jokes at Ars. What’s interesting is that although the AI model aced the task of creating an original song for each topic, both songs start with the same line, “In the depths of the digital domain.” That’s possibly an artifact of whatever hidden prompt Suno is using to instruct ChatGPT when writing the lyrics.

Suno is arguably a fun toy to experiment with and doubtless a milestone in generative AI music tools. But it’s also an achievement tainted by the unresolved ethical issues related to scraping musical work without the artist’s permission. Then there’s the issue of potentially replacing human musicians, which has not been far from the minds of people sharing their own Suno results online. On Monday, AI influencer Ethan Mollick wrote, “I’ve had a song from Suno AI stuck in my head all day. Grim milestone or good one?”

MIT License text becomes viral “sad girl” piano ballad generated by AI Read More »