The Snapdragon 8 Elite is set to be one of the best chips in Qualcomm’s history. Leaks about the formerly known Snapdragon 8 Gen 4 already suggested it, and benchmarks have been confirming it. Over the past few weeks, a few devices powered by the chip have been showing off their might on Geekbench. Now, the Snapdragon 8 Elite-powered realme GT7 Pro has gone through Antutu’s tests, showing massive improvements.

Smartprix got access to the realme GT7 Pro’s results on Antutu and put them side-by-side with those of the iPhone 16 Pro (Apple’s A18 Pro). That said, you should know that Antutu’s tests for Android and iOS devices use different APIs, so the results are not directly comparable. However, you can use them as a reference if you compare them with Antutu results from previous generations of devices.

Snapdragon 8 Elite scores over 3 million points in Antutu; surpasses Apple A18 Pro on Geekbench

The Snapdragon 8 Elite achieved an impressive overall score of 3,025,991 on Antutu. This represents a massive performance boost over the current Snapdragon 8 Gen 3. For reference, the latter scores around 2.1 million points in its best Antutu scores. These results are in line with what we’ve seen so far on other benchmark platforms. For example, in Geekbench, the chip scored 3,260 and 10,051 points on single-core and multi-core tests, respectively.

It’s notable that Geekbench results are comparable between Android and iOS devices. The Apple A18 Pro chip scores around 3,182 and 7,872 points in the single-core and multi-core tests on that platform. In GFXBench, a benchmark focused on graphics performance, the Snapdragon 8 Elite reached 125 FPS and 166 FPS in the DreamWorks Wild Life tests. So, the SoC is ready to take the tech industry by storm as soon as it is available.

The chip could be available in two variants

Leaks about Qualcomm’s upcoming flagship mobile SoC claim that it will use a 2+6 core architecture. The chip would be available in two variants, where the main difference would be the clock speed of the two most powerful cores. The “standard” version would reach 4.09 GHz, while the “Pro” variant would boast 4.32 GHz. The other six cores would have a clock speed of 2.78 GHz in both models.

In addition to the realme GT7 Pro, the iQOO 13, OnePlus 13, and Xiaomi 15 series will be the first devices to boast the Snapdragon 8 Elite. All (or most) of these phones are expected to arrive before 2025.

Apple Pay debuted almost 10 years ago to the day, and Apple is marking the occasion by rolling out some features it revealed at WWDC. If you have iOS 18 or iPadOS 18 installed on your iPhone or iPad, you’ll now be able to use the payment service on third-party browsers on those devices, as well as desktop computers. In the latter case, you’ll be asked to scan a QR code with your iPhone or iPad to complete the payment. However, Apple says the feature is only available in select markets.

It’ll be easier than ever to add supported cards to Apple Wallet on your iPhone. Thanks to the Tap to Provision feature, you can simply tap an eligible NFC-enabled card to the back of your phone, though you may have to enter the security code manually. Again, though, Tap to Provision isn’t available everywhere.

Apple is bringing more third-party buy now, pay later (BNPL) services to Apple Pay. This starts with the ability to check out with Klarna in the US and UK. The company will add more installment payment options in the future, including Citi, Synchrony and eligible Apple Pay issuers via Fiserv in the US. Klarna will also be available as an option in Canada at a later date. Apple discontinued its own Pay Later option earlier this year and it added Affirm as a BNPL option in the US last month.

Meanwhile, US Apple Pay users can redeem rewards on eligible Discover credit cards at checkout. Support for rewards from other issuers and in more countries is on the way.

Update October 18, 8:48AM ET: This story previously indicated that Apple was starting to add third-party BNPL services with the integration of Klarna, but Affirm support was added last month.

If you buy something through a link in this article, we may earn commission.

Modern cybersecurity professionals require advanced technologies to deter, detect and expel hackers, and the predictive benefits of AI can mean the difference between data protection and ruin.

The average cost of a data breach in the U.S. hit a high-water mark of $9.48 million in 2023. Losses have ticked up every year since 2013, even during the global health emergency of Covid-19 when many businesses shuttered. An analysis in IBM’s 2024 data breach report indicates that organizations that employed extensive AI security automation saved $2.22 million, while also lowering cybersecurity insurance.

Industry leaders would be well-served to think about cyberattacks outside the financial implications, as well. Should your organization pay a ransomware demand or right the ship after a crushing malware attack, the reputational damage can far outweigh the dollars. When hackers steal confidential, sensitive and personal identity information, those in your orbit are negatively impacted. Employees, customers and industry partners may file civil actions. 

And, when word gets out that your enterprise cannot protect personal data, business can get eerily quiet. It’s not uncommon for an institution to file for bankruptcy within one year of a significant breach of trust. Fortunately, AI cybersecurity can harden your defenses and make cybercriminals look elsewhere for low-hanging fruit.

What role does AI play in cybersecurity?

There are wide-reaching benefits to integrating AI into an operation’s cybersecurity posture. The lengthy list, which we’ll briefly cover here, does have one central theme — reaction time. The bedrock of the thought leadership behind using AI in the data protection sector involves reducing how long it would otherwise take to detect and expel hackers.

The role AI plays in today’s lightning-quick hacking landscape can determine whether companies suffer stinging losses and hiccups or walk away unscathed. When you consider how fast a sophisticated cybercriminal can work, it’s abundantly clear why time is on the bad guys’ side unless we do something about it.

• Ransomware attacks: These hacks usually take 4 hours, but advanced persistent threats can take over a business network in 45 minutes. Ransomware attacks occur every 11 seconds.
• Phishing emails: Almost 30% of all phishing emails are opened by their recipients. These malware-laced communications account for 91% of all cyberattacks.
• Malware deployment: Hackers deploy malware at a rate of 11.5 attacks per minute.

The average hacker needs only 9.5 hours to pilfer off valuable and sensitive digital assets. Cybercriminals can operate with impunity if no one is monitoring activity while the business is closed and staff are fast asleep. Operations without AI, machine learning (ML) and other advanced technologies typically average 197 days to notice a breach and another 67 days to contain it. Hackers are more than happy to hide in plain sight and copy incoming data until you expel them.

The benefits of using predictive AI technology

The fundamental element of AI in cybersecurity may be its time management effectiveness. It’s important to understand how this forward-looking technology benefits an organization’s overall cyber hygiene. Here are some ways AI delivers quantitative and qualitative data security benefits.

Advanced threat detection

The ability of AI to sift through massive amounts of data seemingly at light speed cannot be matched by human beings. Programmed to learn and identify even subtle anomalies in network traffic, user activity and system logs can make it difficult for hackers to go undetected. Generating a real-time and ongoing analysis of wide-reaching movement, anything that deviates from predictive patterns gets flagged. A cybercriminal or deployed malicious software triggers an immediate threat detection alert. The most skilled perpetrator could not get the 45 minutes needed to effectively insert a ransomware file.

Behavioral analytics

To say that AI exceeds expectations in terms of behavioral analytics would be something of an understatement. ML, largely a sub-category of AI, involves following and understanding consistent patterns. For example, a legitimate network user enters a username, password, then a two-factor authentication code. Once inside the system, staff members carry out relatively consistent tasks. That means they open the same programs, access similar data and perform these duties in a uniform manner.

When a hacker orchestrates an attack, the digital burglar isn’t interested in filing incident reports or tabulating inventory. Cybercriminals head for valuable and confidential information that can be sold on the dark web. Because AI and ML follow the behaviors of users — sometimes down to keyboard strokes — alarms are triggered, and prompt actions are taken to confine and expel the threat.

Reduce fault threat alerst

Before organizations started adopting AI and ML, responding to false alarms seemed like the cost of doing business. That’s largely because the alternative was not knowing when a genuine threat was in progress. In terms of efficiency, pre-AI threat detection was a lot like a fire department responding to dozens of alarms being set off by overly sensitive heat detectors.

The rise of AI has been a game changer in terms of decreasing false alarms and reducing the time managed IT and security officers spend vetting each and every one. As technology adapts to common false positives and learns to distinguish between low-level and heightened irregularities, cybersecurity professionals spend fewer wasted hours.

Non-stop threat monitoring and learning

Although people and most machines require downtime, AI works relentlessly to identify abnormalities. During this never-ending process, technology continues to accumulate actionable information. It can adapt to changes in the digital landscape and be reconfigured to assess new norms. The alternative to AI would be hiring a full-time staff and checking systems activities 24 hours a day, 7 days a week. For many organizations, the cost of non-stop threat monitoring can prove prohibitive.

Getting comfortable with AI automated incident response

One of the processes that AI delivers involves automated threat responses. Not every business director feels comfortable allowing technology to push back on threats, be they malware, ransomware or a human attempting a blunt-force attack. There’s a certain loss of control that accompanies letting the so-called “machines take over.” But automated incident responses may actually be in your best interest.

Industry leaders can choose their comfort level regarding which threats are handled by the technology and which get elevated for a real person’s attention. Low-level threats are typically managed by AI, and it’s commonplace to have AI start the threat containment efforts while security professionals respond to an alert. These rank among the benefits companies gain from automating varying incident responses.

• Speed and efficiency: Pre-determined responses to emerging threats happen immediately. The speed at which AI can address these issues helps efficiently mitigate risk.
• Minimize human error: The majority of successful data breaches can be traced back to human error. Technologies such as AI and others carry out the procedures and duties assigned to them. You can’t trick AI into allowing users to access data deemed off-limits.

Integrating AI and ML may be one of the most cost-effective ways to harden your cybersecurity position. It does the work of dozens of humans faster and more efficiently without logging overtime hours. Adaptable to wide-reaching networks and architectures such as zero trust, its ability to sift through massive amounts of data, identify patterns and constantly learn makes it invaluable in risk management. When a threat actor finds a way into your network or an insider attempts to steal a trade secret, they cannot escape AI’s watchful eye.

John Funk is a creative consultant at SevenAtoms.

Tech billionaire Elon Musk gave a $1 million check to an audience member at a rally in Harrisburg, Pennsylvania, where he’s been Trump-stumping, on Saturday. And he said he will continue to give that prize to one random voter per day until the Presidential election on November 5th. The move could very well break US election law.

Only people who have signed a petition from Musk’s America PAC are eligible for the sweepstakes-style offer. And only registered voters are eligible to sign the petition. According to its terms, Musk will give $1 million to someone in Pittsburgh today, but will later expand to anyone living in swing states like Georgia, Nevada, or Arizona.

It’s against federal law to impel anyone to vote or get registered to vote and Musk’s $1 million prizes are “clearly illegal,” writes UCLA political science professor Rick Hasen in his Election Law Blog. He cites 52 USC 10307 (c), which states that anyone who “pays or offers to pay or accepts payment either for registration to vote or for voting” can potentially see $10,000 in fines or five years in prison. There are also other legally dubious aspects of the pledge, include bonuses for signers and for getting others to sign it.

Pennsylvania Governor Josh Shapiro said on Meet the Press this morning that the offer is “deeply concerning.” He stopped short of calling it illegal, stating that he’s no longer the state’s Attorney General, but said the ploy is “something that law enforcement can take a look at.”

Hasen told NBC News that the fact the petition is only open to registered voters may be what puts it over the top. Federal authorities can either prosecute Musk or warn him to stop the scheme, he said. The US Department of Justice would have to charge him with violating the law in federal court; a federal agency or private party could also file a civil lawsuit.

Either one of these would take time. Musk could be trusting that Trump, who has pardoned allies and is unlikely to let the DOJ punish a major donor, will win and make the legal questions irrelevant. Even if the former President doesn’t win, Musk already has a track record of flouting regulation with few consequences in the US. But conversely, it’s an open question how much Musk’s sweepstakes will help Trump — and whether it can make up for rumored weaknesses in the candidate’s traditional campaign ground game.

Sony‘s camera department has had an unusually quiet year. For the entirety of 2024, we’ve seen just one new Sony mirrorless camera. And, as impressive as the Sony ZV-E10 II is with its 26MP APS-C sensor and 4K video features, the sub $1,000 beginner video camera – that’s an iterative update – isn’t much for Sony fans to get excited about.