A US Navy handler assigned to the Marine Mammal Company's explosive ordnance disposal unit rewards a bottlenose dolphin following a training event. Petty Officer 1st Class Bruce Cummins/US Navy Tuesday's Pentagon briefing took a different turn with a question on "kamikaze dolphins." Pentagon leaders dismissed reports that Iran possesses this capability. Dolphins have long been used by the US military for maritime surveillance and mine detection. During a Tuesday morning press briefing, Pentagon leaders fielded an out-of-the-ordinary query from a reporter seeking clarity on reports of Iran's military using "kamikaze dolphins." "I haven't heard the kamikaze dolphin thing," Gen. Dan Caine, Chairman of the Joint Chiefs of Staff, replied, wondering aloud if the idea was similar to "sharks with laser beams," a wry reference to the early-aughts' Austin Powers movies. "I can't confirm or deny whether we have kamikaze dolphins," Secretary of Defense Pete Hegseth added. "But I can confirm they don't," he said, referring to the Iranians. Following a Wall Street Journal report late last month that mentioned an Iranian threat to use a range of capabilities from submarines to mine-carrying dolphins against US warships, discussions of one-way attack dolphins have been circulating online. Rumors of Iranian "kamikaze dolphins" have been around for over two decades, stemming from reports that Tehran had bought dolphins that had been trained to attack ships and combat divers for the Soviet navy. While the Iranian armed forces may not have dolphins in their arsenal of combat capabilities, the US military has a long history with marine mammals, which are well-suited for tasks like locating underwater mines or alerting handlers to enemy swimmers. Dolphins are exceptionally intelligent animals, with a keen internal sonar system that makes them excellent mammalian sensors. The US has trained bottlenose dolphins to carry out a range of military tasks, from mine detection to harbor defense, and has been using them in a mix of missions since the late 1950s. A handful of these dolphins, for instance, were used briefly in the early 1970s to guard an Army pier in Vietnam, and others were deployed from a US base in Bahrain during the Tanker War in the 1980s to help protect US Navy vessels. The 1996 Republican National Convention in San Diego used dolphins for maritime protection, and then they saw service again seven years later in the Persian Gulf clearing mines for US forces before the invasion of Iraq. In 2015, the Navy's Marine Mammal Program at the Space and Naval Warfare Systems Command in San Diego, California, told Business Insider that the program oversaw training for 85 dolphins and 50 sea lions. It's unclear how many are part of the program today. While the US military has widely acknowledged the value of dolphins as a military working animal for key detection and protection missions, the Navy has challenged reports stating that dolphins have been trained to kill enemy swimmers. A 1990 New York Times article regarding the US Navy's decision to halt part of its dolphin program at the time cited former trainers who said that some mammals were being trained to kill with nose-mounted guns and explosives. A Navy spokesperson said at that time that dolphins were only used for surveillance, detecting objects, location, marking, and recovery. "We don't train animals to kill people," the spokesperson said. There have been other such assertions since the Times article. Famous ocean explorer Robert Ballard, who found the Titanic, wrote in his 2021 memoir that he was approached about "training dolphins to kill enemy divers in Vietnam." He said that he declined because "it didn't feel right to put the animals in that position." And a former special operator told Business Insider in 2023 that dolphins have played defense against forces like the Navy SEALs in training, attempting to "kill" or capture the "enemy" combat swimmers, preparing them for a potential encounter with an adversary dolphin or other militarized animal. In 2022, the Navy considered shuttering the Marine Mammal Systems program and ending the mission of using dolphins and sea lions to help find and neutralize underwater mines, with plans to instead rely increasingly on advanced sensors and underwater vehicles. However, it was determined that the technology had not yet advanced beyond the dolphins' capabilities. The Navy program is still active today, focused on detection and overseen by the Reconnaissance and Interdiction Division of Naval Information Warfare Center Pacific. Read the original article on Business Insider

State says chatbot claimed to practice medicine, gave invalid license number.

With Apple's latest operating system updates, users will reportedly have their pick of which third-party AI models they want to use for a host of tasks.

Building with a smaller team has its benefits and challenges. skynesher/Getty Images Coinbase plans to lay off 14% of staff, shifting focus to AI efficiency and smaller teams. Three leaders running AI-driven teams of fewer than 10 people share key lessons from operating lean. They say speed is the biggest advantage of AI, but it can also create new challenges. More and more companies are reducing their staff and leaning into AI. Coinbase CEO Brian Armstrong is the latest tech exec to head in that direction. Armstrong said the fintech company plans to lay off around 14% of its workforce. He shared his message publicly on X on Tuesday morning, saying, "The pace of what's possible with a small, focused team has changed dramatically, and it's accelerating every day." The appeal is simple: smaller teams mean faster decisions, lower costs, and more control. But as AI changes the number of people companies actually need, new challenges have emerged. Business Insider has spoken with over a dozen executives at lean AI startups with fewer than 10 full-time employees over the past few months as part of our Tiny Teams series. They've all agreed that speed is the biggest advantage that AI has given them. However, with that speed, they've cited having to work harder to guide creativity, hire the right job candidates, and minimize potential mistakes. Below are three founders who shared the biggest upsides and challenges that AI has presented them and their tiny teams. Their words have been edited for length and clarity. Business Insider wants to talk to founders of AI-powered companies with fewer than 10 workers, and employees working alongside AI agents and bots, to understand what the "Tiny Teams" era really feels like — the wins, the fears, and the human skills that stand out. Share your story by filling out this quick form. The hardest part of a small team is balancing quick executions and meaningful creativity Nathaneo Johnson is a 22-year-old CEO of Series, an AI social network, based in Chelsea, New York. A lot of CEOs who are running older companies are in a relatively tough position to have to lay off a significant number of employees during the awakening of AI automation. The AI era is all about speed, which makes lean specialized teams thrive, but the hardest part, day to day, is trying to restrict meetings without losing creative brainstorming time. I started Series with my cofounder during this exact era, so there wasn't a need to hire at scale when there are AI agents for practically anything. However, creativity and vision will always be where humans win. On a tiny team, it's always a game of whether we encourage creativity and collaboration or push more heads-down execution. We could be the company that never talks and just works in Slack all day, and that seems to optimize productivity. The flip side is that there could be a brainstorming meeting that turns into a 10x idea, and we don't want to lose that. In the next five years, you won't need someone to be a heads-down specialist; you'll need someone who's creative or visionary. We can stay profitable when we're lean, but finding the right hires is becoming more challenging Sidhant Bendre is a 26-year-old cofounder of Oleve, an AI-driven consumer software portfolio company, based in New York. Keeping our team small stemmed from a desire to be profitable. The fastest path to growth for us was to figure out tricks and hacks to do more with AI and fewer people. But on tiny teams, there's no middle management layer to catch sloppiness, and there's no room for people who aren't thinking about how their work affects what comes next. When everyone is a specialist, moving fast with a real sense of agency, one bad result doesn't just mean one bad result; it compounds into an entirely bad system. That's why the bar has been raised for each individual contributor, making hiring challenging. We've had to create more involved engineering recruitment processes because we need to vet potential employees more closely. We've seen take-home tasks from job applicants where someone clearly just fed a prompt into ChatGPT and submitted whatever came back without critical thought. It's a skill to use AI to fully drive results, but that can be temporary when someone lacks the deeper knowledge needed to fix issues that arise. The faster we move, the riskier it is for junior employees to make mistakes Charles Swann is a 44-year-old founder of an AI startup based in Boulder, CO. I founded my startup around two years ago, and I have one full-time employee. My business is in the marketing technology space, and I needed someone with a deep connection to modern culture and social media. So, I hired a 24-year-old from my neighborhood to be my growth and brand specialist. I don't need a huge team of founding engineers, each with a six-figure salary, to launch a product. I can use AI to have my junior employee produce senior-level work. AI, specifically Gemini, serves as a middle management layer, helping her level up the work she produces. But, there's always a risk in relying on AI to teach my employee how to gain years of experience in seconds. Hallucinations and feedback loops can occur when someone lacks the experience to know when to redirect. What I've done to help safeguard us is create a collection of prompt starters that I can copy and paste into the chat to save time and help keep the AI focused on relevant context. We might encounter potential mistakes or hallucinations, but I'd rather have those mistakes come up and have to course-correct than not be able to move at the pace we are. Do you have a Tiny Teams story to share? Contact this reporter, Agnes Applegate, at aapplegate@businessinsider.com. Read the original article on Business Insider

Info is scant, but such breaches can reveal where a security product's controls are located and how detections are designed, giving attackers a leg up.

GameStop CEO Ryan Cohen appeared on Fox Business's "Making Money with Charles Payne." Screenshot via Fox Business GameStop CEO Ryan Cohen smiled and laughed through his interview with Fox Business's Charles Payne. Cohen was more high energy and generally less awkward and combative compared to his viral CNBC interview the day prior. Payne was borderline fawning, comparing Cohen to a quarterback and telling him he remained a shareholder. If GameStop CEO's awkward CNBC interview was an away game, his Fox Business appearance the following day felt like a home game. Ryan Cohen seemed happier to be on TV during Tuesday's edition of Fox Business' "Making Money with Charles Payne," where the CEO again discussed GameStop's unsolicited bid to acquire eBay. It was his first televised interview since his bizarre segment on CNBC's "Squawk Box," during which his repetitive answers and stated confusion at their questions elicited exhalations and chuckles from the hosts. This time around, Cohen was all smiles. He said it was "good to see you by the way, Charles" and laughed several times throughout his appearance. (The CNBC hosts got a "good morning" and a brief smile that faded into a dead-pan expression, which remained for much of the interview.) With a friendly host, GameStop CEO Ryan Cohen was all smiles. Screenshot via Fox Business It may have helped that "Making Money" is filmed in the afternoon, unlike "Squawk Box," which is early morning. It also may have helped that Payne was borderline fawning, saying his own opinion was that Cohen was the "key to this," comparing it to the New York Giants trading for "a better quarterback" and that he found traditional Wall Street's skepticism of the proposed deal "frustrating." "They paid themselves over the last few years, gobs, millions and millions of dollars," Payne said in reference to eBay's board. "They never buy back their own stock. They've done nothing really innovative at all with respect to that. And here, I think, really, for me, that my opinion is you're the linchpin, you're the key to this." Cohen made a greater attempt to explain the proposal to buy eBay, which is worth several times what GameStop is today. When the CNBC hosts asked how he would fund the deal, he repeatedly told them to look at GameStop's website. In comparison, for Fox Business, he talked at length about the possibility of "dilution that is accretive to earnings per share" and outlined a vision for what he could make more efficient in eBay's business. There were still some similarities between the two TV appearances. Cohen spent much of the first portion of the interview looking down while answering — which he did throughout the CNBC spot — but later spoke directly to Payne and generally seemed higher energy. He also took another jab at those he said predicted GameStop's failure. On CNBC, it was directly to the anchors as "you guys"; on Fox, it was aimed at "experts" who have "never run a business." On "Making Money," Cohen took more time to explain the deal, even becoming expressive with his hands. Screenshot via Fox Business There were also non-answers. These were given with more smiles than CNBC's, though. At one point, Payne asked for his thoughts about a Semafor story that said Cohen "squandered what goodwill he had" with the awkward interview. There was a long silence, followed by Cohen's "um." Then, the CEO laughed. "Can't take it too seriously," he said. "It's OK." Earlier in the interview, Cohen repeated a line that he used on CNBC. This time, with more energy in his voice. "We'll see what happens," he said. Payne then segued to Cohen's view that "American capitalism is rotting from the head down." "Let's talk about that, because I tend to agree with you," the host said. As the interview wrapped, Payne told Cohen he was wishing him luck: "I'm still a shareholder, and I still have faith." Read the original article on Business Insider

Larry Fink runs the world's largest asset manager, BlackRock Anna Moneymaker/Getty Images BlackRock CEO Larry Fink said at Milken that his firm is partnering with a hyperscaler. The partnership, intended to build data centers, will be announced later this week, he said. Fink has positioned $13.9 trillion BlackRock to fund the buildout of infrastructure for AI. BlackRock CEO Larry Fink is one of the most bullish people in the world about artificial intelligence, and he just announced another AI bet. Speaking at the Milken conference in Los Angeles on Tuesday, Fink said his firm will announce a partnership with an unspecified hyperscaler later this week to build data centers. "The future is today," Fink said. BlackRock did not immediately return requests for comment. He has positioned his $13.9 trillion asset manager to provide the financing for the technology's buildout, with a focus on infrastructure such as data centers and energy investments. BlackRock's $12.5 billion acquisition of Global Infrastructure Partners, a private market asset manager, in late 2024 was just one part of Fink's broader ambitions. BlackRock has teamed with Microsoft, Nvidia, Abu Dhabi-backed investment fund MGX, and others to invest in data centers. The group, led by GIP, announced plans to buy Aligned Data Centers in a deal valued at about $40 billion last fall. Other alternative asset managers are also pouring billions into AI infrastructure, even working together more often on deals rather than strictly competing. They're also partnering on how to get more of their portfolio companies to use the tech. On Monday, private equity firm Blackstone and other managers announced a joint venture with Anthropic to ultimately create a playbook for companies to use the tech. For their part, the four biggest hyperscalers, Amazon, Microsoft, Meta, and Google, have planned $725 billion in spending this year, which has some worried about the sheer size of the bet. As Fink told the Milken audience, despite the astronomical funding figures announced every day by tech companies and AI-obsessed asset managers alike, "I ultimately think we are going to see a shortage of capital" to build out the needed infrastructure. As Brookfield CEO Bruce Flatt said on the same Milken panel, "We're rewiring the world." "We're short power, we're short compute, we're short chips," Fink said, adding that he thinks there will one day be a new asset class of compute futures, to help companies hedge against a jump in compute costs like agriculture or energy companies do with certain crops or fuel sources. "I don't believe we're moving fast enough," he said. "There is not an AI bubble. There is the opposite." Read the original article on Business Insider

Discussion | Link

Christophe Fouquet, who became ASML's CEO in 2024 after more than a decade at the company, sat down with this editor on the rooftop deck of his Beverly Hills hotel Tuesday morning ahead of his appearance at the Milken Institute Global Conference. Dressed in a blue suit and white shirt, he was relaxed — even when the conversation turned to the rivals.

The UC Berkeley Center for Long-Term Cybersecurity (CLTC) offers tools and support to schools, local governments, and non-profits as they defend themselves against a growing volume of cyberattacks.

Greg Brockman arriving in court on Monday Bloomberg/Bloomberg via Getty Images OpenAI cofounder Greg Brockman returned for a second day of testimony at the Musk v Altman trial. He recalled a "celebratory" gathering where Musk's then-girlfriend, Amber Heard, served whiskey. Musk was excited by OpenAI's for-profit prospects — then resigned in a fight over control, he said. In the summer of 2017, Elon Musk hosted a festive gathering for OpenAI's cofounders at his Bay Area "haunted mansion." Amber Heard, his then-girlfriend, poured the whiskey as they celebrated the company's future. Days later, as the cofounders gathered again, Musk erupted in a bitter play for control of the company. "He stood up, and he kind of stormed around the table," OpenAI cofounder Greg Brockman told jurors on Tuesday in the ongoing federal trial pitting Musk against OpenAI CEO Sam Altman. In detail-rich testimony, Brockman described Jekyll and Hyde-level swings of temperament by the Tesla CEO — from happy party host one day to tantrum-throwing control freak the next. The testimony by Brockman, a longtime Altman loyalist who is now OpenAI's president, was not only vivid but could prove strategically helpful for Altman's side. In taking Altman to trial, Musk hopes to prove that Altman and Brockman deceived him into donating $38 million to OpenAI when it was a nonprofit dedicated to helping humanity, only to pivot to a for-profit for personal gain. But Brockman told jurors Tuesday that throughout 2017, Musk eagerly pursued turning the company into a for-profit. Musk felt that keeping OpenAI solely nonprofit was deterring investors, including Bill Gates, Brockman testified. "He asked Gates four times to donate, and Gates didn't even come to the office," Brockman told the jury. Ultimately — at that second cofounder meeting, days after the party where Heard poured the group "some very good whiskey" — Musk played his big card, cash. "It started out very pleasant," Brockman said. "I'd like you to have this Tesla painting," as a gesture of goodwill, Brockman recalled cofounder Ilya Sutskever telling Musk of a painting he had commissioned. "Then the conversation turned to equity," Brockman said. "And something really changed. Something just shifted in him," Brockman said of Musk. "And he was angry. He was upset." Musk wanted a controlling, 51% equity stake in OpenAI — and to be its CEO, Brockman told the jury. "He said he deserved more because he had started the most multi-billion-dollar companies in history, that he had zero failures," Brockman testified. "Look, you guys are great," Brockman said Musk continued. "But I can start another AI company tomorrow, like in one tweet." This was something the rest of the founders — Brockman, Sutskever, and Altman — objected to. Musk "knows rockets" — not AI, Brockman said "Look, he knows rockets. He knows electric cars. He did not, and I believe does not, know AI. That was a major concern," Brockman told jurors. "And Ilya and I did not think that he was going to spend the time required to actually get good at it." "At the end of the meeting, he just sat quiet for several minutes, just thinking," Brockman told jurors. "And he said, 'I decline.' He was not going to accept these terms." Musk got up, "kind of stormed around the table," and lunged toward him, Brockman testified. "I thought he was going to hit me," Brockman told the jury. Instead, Musk grabbed the painting of a Tesla that was behind Brockman, he testified. Musk's final words, painting in hand, as he "started to storm out of the room," were these, Brockman said: "I will withhold funding until you decide what you're going to do." "And then he left the August 29 meeting." Then-board member Shivon Zilis called later that night, Brockman testified, saying she had spoken with Musk and "Hey, it's not over." But "it is wrong to give unilateral control," Brockman said of opposing Musk's terms. Later Tuesday, a lawyer for Musk grilled Brockman on emails from the months before his split with OpenAI. In the emails, Musk continued to push for the company to remain primarily nonprofit. After Musk's departure, "you transferred all of the intellectual property from the nonprofit to the for-profit, right?" the lawyer, Stephen Molo asked. Musk argues in his lawsuit that the nonprofit was gutted and sold off in order to turn OpenAI into a "market paralyzing gorgon." "You transferred the intellectual property for a reason, right?" Molo asked Brockman. "I mean if it was worthless, you would have just left it there, right?" "I think, yes — I'm not sure, exactly," Brockman responded. "I'm not certain how it was implemented," he added. Brockman also told jurors that Musk was pivotal to OpenAI's founding. "He provided the predominant funding. He provided the vision," Brockman said. "So I'm deeply grateful for his contributions." Read the original article on Business Insider

Daemon Tools users: It's time to check your machines for stealthy infections, stat.

Hannah Murphy / Financial Times: Sources: Meta is building agentic tools, including an OpenClaw-like assistant powered by its new Muse Spark AI model to help users create AI bots  —  Social media platform invests in equivalent to OpenClaw that aims to seamlessly carry out everyday tasks for users

Kennedy has made—and continues to make—many false claims about SSRIs.

Rachel Metz / Bloomberg: Musk v. Altman: on his second day of testimony, Greg Brockman said that OpenAI expects to spend $50B on computing in 2026, up from $30M in 2017  —  OpenAI expects to spend $50 billion on computing power this year to support its artificial intelligence software, according to co-founder and President Greg Brockman.

Databricks’ monitoring infrastructure has more than tripled in size over the last year, ...

Bloomberg: Musk v. Altman: Greg Brockman testified about tense negotiations with Musk in 2017, saying “he knows rockets, he knows electric cars” but “does not know AI”  —  OpenAI President Greg Brockman testified that Elon Musk called a ChatGPT predecessor “stupid,” …

Kyt Dotson / SiliconANGLE: Subquadratic launches with a $29M seed and debuts SubQ, an LLM that uses a subquadratic sparse attention architecture to achieve a 12M-token context window  —  Subquadratic, a company developing a novel generative artificial intelligence model, launched today with $29 million in seed funding.

Discussion | Link

Mary Ann Azevedo / Crunchbase News: Blitzy, which says its platform autonomously completes months of software development, raised $200M at a $1.4B valuation, bringing its total funding to $232M+  —  Blitzy, an autonomous software development startup, said it has raised $200 million in a funding round that values it at $1.4 billion …