Start Up No.2786: old Bitly links being hijacked, Microsoft explains outage, US tries to block AI chips to China (again), and more


Don’t get too invested in the idea some are pushing of Jensen Huang being reinstated at Nvidia – the board is against it. CC-licensed photo by Village Global on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.


A selection of 9 links for you. Unseated. I’m @charlesarthur on Twitter. On Threads: charles_arthur. On Mastodon: https://newsie.social/@charlesarthur. On Bluesky: @charlesarthur.bsky.social. Observations and links welcome.


Political and corporate X posts have links redirecting to vulgar, racist or random websites • CNN Business

Auzinea Bacon:

»

Links dating as far back as 2009 from URL-shortening service Bitly were affected, and only corporate and political X accounts appear to have been targeted, with President Donald Trump’s personal account seemingly hit the hardest. At least a dozen of Trump’s posts on X now direct to various inappropriate and unrelated sites.

Public figures — like Microsoft co-founder Bill Gates, former Canadian Prime Minister Justin Trudeau and former Vice President Mike Pence — as well as fast-food chains McDonald’s and Burger King, were also hit to varying degrees.

Bitly URLs are shortened while holding the original and longer link. Bitly links do not automatically expire, according to the company’s website, but paid users can redirect links.

Bitly said Sunday afternoon in a post on X that it had “identified an issue that allowed bad actors to repurpose certain old, inactive links to redirect people to content unrelated to the original posts.” The company said it suspended the accounts that altered links and were working to disable the affected URLs.

“Our Trust and Safety team is working to address the issue and strengthen protections against similar misuse in the future,” Bitly wrote. Both Bitly and X have not yet responded to CNN’s request for comment.

It’s unclear whether hackers made custom Bitly links, which effectively allow them to copy the original URL and redirect to any website. That may have been possible if the original Bitly link to the website no longer exists or has expired.

Bitly links may be susceptible to artificial intelligence scans and exploiting mechanisms to redirect links, noted Ari Lightman, a digital media professor at Carnegie Mellon University’s Heinz College of Information Systems and Public Policy. The redirected links seem confined to posts made between 2009 and 2014.

«

The redirects were often darkly humorous, which suggests (to me anyway) that they were the work of script kiddies who discovered a weakness. Though no doubt we’ll now discover it was a careful plot by Russia whose reason we just haven’t yet fathomed.

The June 2016 XKCD “Theft Quadrants” is probably relevant here.
unique link to this extract


Microsoft identifies months-old bug as a root cause of global outage • The Guardian

Johana Bhuiyan:

»

The office software giant issued an initial report on Monday that identified three factors that led to the 24-hour blackout of its services: a months-old bug, malfunctioning internal connections , and server capacity. Each of these factors may have easily been resolved on their own, but, when taken together, they led to a major outage that was harder to mitigate, Microsoft executives said.

The disruption began Thursday with US users reporting interrupted access to their Outlook email accounts. As it continued into the next day, millions of people worldwide said they could not access their email in any capacity as well as other Office 365 services. Jails could not release prisoners, hospitals turned to paper records, schools lost contact with parents, and police departments could not receive crime reports. Microsoft said it had resolved the issue and restored service by Saturday, also issuing an apology. The company did not attribute the blackout to a cybersecurity breach.

In a post on X on Monday, Charles Lamanna, the executive vice-president of Copilot, Agents, and Platform at Microsoft, said there was a latent bug or defect from July that caused Microsoft’s servers to “become overwhelmed and unresponsive”.

Lamanna wrote that the company typically has two available networking gateways, or entry and exit points that connect separate systems and allow data to pass between them. However, one of them was not working for Microsoft Teams which overwhelmed the primary gateway. Subsequently, over 14% of the company’s server capacity was also not working due to an unrelated issue from the previous day.

“Taken together, the increased load and the reduced server capacity, caused the network gateway to encounter the latent bug from July,” Lamanna wrote. “This caused requests to build up until servers ultimately stopped responding. This occurred during the peak window of traffic in the United States.”

The company also had trouble mitigating the bug because it originated months before, Lamanna said, so it would not be easy as rolling back a recent change to undo the defect.

“Recovery took longer than usual because as servers returned to service, they were immediately overwhelmed by waiting traffic and became unresponsive again,” he wrote.

«

The cascade effect is well known as a cause of system collapse, and as a reason why a stopped system becomes impossible to get up again, because all of the required capacity has to be available at once. But throwing every switch at once is nearly impossible. If you’re Microsoft, still challenging.
unique link to this extract


Major US banks report widespread service outages • CBS News via MSN

Megan Cerullo:

»

Customers at Bank of America, JPMorgan Chase, U.S. Bank and other major financial institutions reported service outages on Monday, according to Downdetector.com. 

Other banks seeing a spike in customer complaints about difficulty accessing their accounts include Capital One, Truist and Wells Fargo, while users also reported problems with the digital payments app Zelle, according to the outage tracker. 

However, Zelle told CBS News that its service was “operating as it should,” but that some users could be experiencing issues related to problems at their financial institutions, which could be delaying Zelle payments. 

“Consumers use Zelle through participating banks and credit unions, which manage their accounts and hold their funds,” a Zelle spokesperson said. “Consumers do not have a separate Zelle account, and Zelle does not hold consumer funds.”

Zelle’s network operator, a company called Early Warning Services, is owned by financial institutions including Bank of America, Capital One, JPMorgan Chase, PNC Bank, Truist, U.S. Bank and Wells Fargo.

Chase Bank’s services are experiencing a technical bug issue, not a cyber issue, a spokesperson said Monday.

“Due to a technical issue, some transactions are taking longer than usual to complete but customers still have access to their accounts and funds,” a Chase spokesperson said in a statement. “We are working to resolve this as quickly as possible.”

«

Microsoft, aircraft, banks.. not to mention the freezers the other week. So what’s next?
unique link to this extract


US moves to cut China’s remote access to AI chips • Ars Technica

Ashley Belanger:

»

On the heels of the Nvidia scandal, the Trump administration is preparing to crack down on foreign data centers that are fueling China’s advancements in artificial intelligence by allowing remote access to otherwise restricted high-end chips necessary for AI training.

In a summary of the rule change, the Commerce Department said it was closing a loophole that previously allowed foreign rivals like China, Russia, and Iran to bypass export curbs on certain advanced computer chips by instead renting access to those chips from data centers in Indonesia, Malaysia, Singapore, Thailand, and Brazil.

US rivals could possibly use that access to train and run advanced AI models “capable of supporting military- and intelligence-related activities that could threaten US national security or foreign policy interests,” the Commerce Department said. Among threats, the US fears rivals could use advanced AI to develop “more effective weapons of mass destruction and advanced conventional weapons, coordinated cyber-espionage and targeted cyber-disruption, algorithmic theft, and weaponized disinformation.”

Now, the rules have been tightened up so much that the South China Morning Post reported that, “as a result, any advanced chips made for a Chinese firm are assumed to be restricted, and factories must ask the US government for a license before shipping them.”

The rule change takes effect 90 days after it’s posted to the Federal Register. Moving forward, companies headquartered in any of those countries deemed to “present a higher-than-normal risk of diversion” to China, Russia, and Iran must apply for a license with strict know-your-customer controls when transferring high volumes of advanced chips. If a license is granted, companies will have to “perform robust due diligence,” submitting monthly reports on transactions that thoroughly detail all parties involved, as well as their parent companies, and the final destination for chips sold.

«

Let’s see how quickly China figures out a way to route around this. Probably not 90 days.
unique link to this extract


Nvidia board not considering reinstating Jensen Huang as chief executive • Financial Times

Michael Acton, Stefania Palma, Abigail Hauslohner, Aime Williams and Desert Sevastopulo:

»

Nvidia’s board is not considering reinstating Jensen Huang as the company’s chief executive, according to people familiar with the matter, despite Donald Trump saying he was open to such a move after the tech boss was ousted over an infidelity scandal.

The news marks the latest twist in the dramatic saga that has unfolded since it emerged that Huang had been engaged in an extramarital affair with a woman, Alice Huang, whom US intelligence officials believe was recruited to spy by China’s intelligence agency.

The Nvidia chief resigned last week as the $4tn company’s board engaged in an independent probe into the spectacular leadership crisis. But a campaign among some leading tech figures to bring back Huang had gathered steam, while the US president said on Monday he “wouldn’t mind” if he were reinstated.

The FT reported on Sunday that Huang had discussed with confidants the possibility of returning to Nvidia.

The chipmaker’s board, which includes a number of longtime Huang associates, is carrying out an independent inquiry into his conduct. The board has said it will consider both internal and external candidates for chief executive.

Last week, the FT reported that the board had approached Arm chief executive Rene Haas in its search for a successor to Huang.

«

There’s a mad bunch of people who want the board to Do An OpenAI and put him back in place, as they did with Sam Altman’s brief exit. But they don’t control Nvidia, unlike those who pushed for Altman to be reinstated.
unique link to this extract


IHeartMedia denies plans to sell podcast business • Hollywood Reporter

Caitlin Huston:

»

iHeartMedia is denying reports that the company is looking to sell its podcast business.

On Wednesday, Bloomberg reported that the audio company has been in discussion with investment bank Moelis & Co and others to possibly sell part of all of its podcast business, which is one of the biggest podcast networks, with shows including Stuff You Should Know, Las Culturistas, The Breakfast Club and more. The sale was positioned as a means for the company, which is the largest US radio station owner, to raise cash and boost its stock price.

A spokesperson for iHeartMedia disputed the account.

“While we don’t generally comment on rumors or speculation, I can tell you that we have no plan to sell our podcasting business. iHeart has built the number one podcast publisher in America, with unmatched scale, premium talent and strong financial performance. Podcasting is a core strategic growth engine for our company, and we continue to invest in expanding its reach and potential, including through video podcasting,” the spokesperson said.

The report came as iHeartMedia’s share price has dropped 65% in the past five days. Share prices spiked 16% Wednesday after the Bloomberg report.

Podcasting has been a highlight for the company, reporting $174m in revenue for the past quarter, up 24.6% compared to prior year and above guidance. The iHeartPodcast network was the third by weekly reach in the US, according to the latest Edison research.

In comparison, iHeart’s overall business missed expectations for adjusted earnings before interest, taxes, depreciation and amortization, with the number falling 17% from a year ago to $170.5m, with lower than expected guidance going forward.

«

You’d have to be mad to sell the part of the business that is actually working to shore up the part that’s doing badly and is probably facing a long slow decline. What probably really happened is that iHeartMedia is looking to sell the radio business. Podcasting seems like it’s got plenty of life yet.
unique link to this extract


The financial bubble cycles are getting worse • The New York Times

Zachary Carter:

»

It wasn’t too long ago that Silicon Valley was celebrating the greatest invention in economic history. A dazzling new medium called cryptocurrency had arrived — at once an extraordinary investment, a government-proof digital money and perhaps a way to bring world peace. It would end global poverty and cure inflation. The possibilities were endless — until the public face of the crypto revolution landed in federal prison and the American president began winning his family millions through Trump-branded crypto schemes.

Cryptocurrency still exists, though few take seriously the grandiose promises of yesteryear. As the public’s enthusiasm for digital currency faded, some of its most prominent boosters began showering praise on another shiny object: The artificial intelligence universe is just lousy with crypto players. AI “is our alchemy, our Philosopher’s Stone — we are literally making sand think,” trumpeted Marc Andreessen, whose firm has been one of the most influential crypto and AI players over the past decade.

Of course, the extreme prognostications about artificial intelligence aren’t exactly like the woo-woo crypto hype of 2020. AI already has some genuinely useful applications in software development. It might even blow up the world! Maybe that’s why Silicon Valley’s hype machine has been so successful at turning the economy into a massive one-way bet on delivering a still undefined type of “superintelligence” to the globe. Last year, according to the St. Louis Fed, nearly 40% of American economic growth came from the AI build-out.

These days, however, excitement is mutating into anxiety. A major AI-themed hedge fund collapsed over the summer. Days ago, shares of the AI chip giant Nvidia plummeted upon the abrupt resignation of its megawatt chief executive amid allegations of extramarital affairs with a Chinese spy and two women who reported to him. Whatever technological changes may lie in store, the bigger picture is starting to look a lot like a movie we’ve seen before — wild promises and warnings, unusual financing arrangements and the collapse of highflying adventurers.

«

(Gift link.)
unique link to this extract


The Pentagon ignored the Army’s warnings about Iran’s deadly drones • The Atlantic

Simon Shuster:

»

rom a distance, the Iranian drone resembled an arrowhead, pointing downward as it dove out of the sky. Its propeller, powered by the kind of motor typically used in dirt bikes and lawnmowers, produced a high-pitched wail. It struck midmorning on March 1, the second day of the US-Israeli war against Iran, hitting a logistics centre in Kuwait used by the American military. Dozens of personnel from the Army Reserve were working at an office building there.

The incoming drone, a variant of the Shahed, slammed through the roof of the building and detonated its payload. Six American service members were killed and more than two dozen wounded, several of them critically—the deadliest attack on US troops in the war. It will be hard for the survivors to forget the drone’s distinctive noise.

In Ukraine, which has faced years of attacks from this type of drone, people often compare its sound as it falls out of the sky to a Vespa revving its engine. But the noise no longer terrifies Ukrainians the way it once did. Last year, Ukraine’s arms industry developed a cheap and effective way to shoot down Shaheds. The US military knew about this technology months before the war in Iran began. A group of senior American officers, worried about exactly the type of attack that killed those six soldiers in Kuwait, told me they repeatedly urged commanders in the Middle East to deploy the new Ukrainian technology.

The Pentagon ignored their warnings, which included detailed reports sent up the chain of command. Several of the officers who tried to raise the alarm about Iranian Shaheds have since been fired or forced into early retirement.

…The warnings were based on Ukraine’s experience with Russia, which acquired Shahed drones from Iran in 2022. Waves of them have pounded Ukrainian cities ever since, often in swarms of a few hundred per night. Last summer, the armed forces of Ukraine deployed a new type of interceptor to destroy Shaheds in midair. U.S. Army officers stationed in Germany began training on these interceptors that fall, and they asked their Ukrainian allies to produce a report about their new invention.

The document, dated December 2025, was titled in the military’s turgid style: “Operational Experience in the Employment of UAV Interceptors in Countering Shahed-Type Aerial Threats.” It described in detail how Ukraine had developed these defensive weapons. A drawing of an interceptor appeared on the cover of the report, alongside a Shahed with a bullseye on its belly. It was shared inside the military and with some US allies in December, roughly two months before the war in Iran started.

«

(Gift link.)
unique link to this extract


McDonald’s AI-generated mural of Blackburn riddled with errors • BBC News

Yunus Mulla and Stewart Whittingham:

»

Fast food giant McDonald’s has promised “a little more accuracy” after an AI-generated mural with “glaring errors” went up in one of its restaurants.
Staff at the fast food outlet in Blackburn created the poster with AI to highlight some of the Lancashire town’s sights and claims to fame.

However, the image included a cathedral from Liverpool, London’s Royal Albert Hall and a sign proclaiming Blackburn Rovers as 1995 Premier League “crampions”.

Apologising for the errors, the firm said it would commission a local artist to create a new mural at its King William Street restaurant.

«

Bit by bit, we will win the war against the machines.
unique link to this extract


• Why do social networks drive us a little mad?
• Why does angry content seem to dominate what we see?
• How much of a role do algorithms play in affecting what we see and do online?
• What can we do about it?
• Did Facebook have any inkling of what was coming in Myanmar in 2016?

Read Social Warming, my latest book, and find answers – and more.


Errata, corrigenda and ai no corrida: none notified

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.