Techin' Geek

On Friday, Microsoft revealed a set of adjustments aimed at enhancing the quality of its Windows 11 operating system, which significantly involves reducing the number of access points to its AI assistant, Copilot.

The corporation stated it will decrease Copilot AI integrations in various applications, beginning with Photos, Widgets, Notepad, and the Snipping Tool.

Under the theme of “implementing AI where it matters most,” Pavan Davuluri, EVP of Windows and Devices, shared on the company’s blog that Microsoft is taking a more deliberate approach regarding “how and where Copilot integrates within Windows.” He clarified that the objective is to concentrate on AI functionalities that are “truly beneficial.”

This “less-is-more” strategy for embedding AI into current platforms might illustrate the increasing consumer resistance against AI excess. While a significant number of individuals today recognize AI as a valuable resource, there are also worries surrounding trust and safety. For example, a Pew Research study released this month indicated that half of U.S. adults are now more apprehensive than enthusiastic about AI as of June 2025, a rise from 37% in 2021.

This isn’t the first instance where Microsoft has reconsidered its Copilot integrations. Earlier this month, the news platform Windows Central reported that the company’s plan to roll out Copilot-branded AI features across Windows 11 had been discreetly abandoned. This, according to the site, encompassed several system-level integrations within the Settings app, File Explorer, and other areas.

Prior to this, Microsoft had postponed the debut of its AI-driven memory feature, Windows Recall for Copilot + PCs, for over a year as it sought to resolve users’ privacy issues. The Recall feature went live last April, but security flaws are still being uncovered.

Clearly, user input is shaping Microsoft’s actions regarding AI on Windows. Davuluri mentioned that he and his team have dedicated the last few months to listening to the community on how they wish to see Windows enhanced.

The Copilot adjustment is merely one of the modifications underway.

The organization also announced that it will enable the taskbar to be positioned at the top or sides of the screen, provide users with increased control over system updates, hasten File Explorer, enhance the Widgets experience, refresh the Feedback Hub, and simplify navigation within its Windows Insider Program—a community that shares feedback about the future of Windows.

CEO Jensen Huang took the platform at Nvidia’s GTC conference this week in his notable leather jacket to present a two-and-a-half-hour keynote, forecasting $1 trillion in AI chip sales by 2027, asserting that every firm requires an “OpenClaw strategy,” and wrapping up with a wandering Olaf robot that needed its mic turned off. The message was clear: Nvidia aims to be integral to everything, ranging from AI training to self-driving cars to Disney parks.

In this episode of TechCrunch’s Equity podcast, Kirsten Korosec, Anthony Ha, and Sean O’Kane dissect what Nvidia’s expanding network of AI infrastructure alliances genuinely conveys for startups, and delve into more of the week’s news.

Subscribe to Equity on YouTube, Apple Podcasts, Overcast, Spotify, and various platforms. You can also follow Equity on X and Threads, at @EquityPod.

Kalshi is experiencing a challenging week. On Tuesday, Arizona’s attorney general submitted a 20-count criminal complaint against the online prediction marketplace, alleging it operates an illegal gambling enterprise. Now, another southwestern state has taken significant action against the company: A Nevada judge has temporarily prohibited the service from functioning there due to an ongoing court case initiated by state regulators.

In February, Nevada, with the support of its Gaming Control Board, initiated legal action against Kalshi in efforts to prevent the prediction site from functioning. Officials assert that Kalshi has not obtained the necessary gaming licenses corresponding to the type of betting activities its users participate in and that allowing users under the age of 21 to access its services contravenes state law.

Earlier this month, the state sought a temporary restraining order against Kalshi as part of its continuing case. In a state court on Friday, Judge Jason D. Woodbury granted the state’s request and set a hearing for the restraining order for early next month, according to court documents.

In his ruling, Woodbury indicated that Kalshi lacked licensing under the Nevada Gaming Control Act and that, given Kalshi’s practice of taking a commission from contracts acquired through its system, it was evidently engaging in a “percentage game” (which the state classifies as gambling).

Kalshi contends that its registration with the Commodity Futures Trading Commission places it under that federal agency’s exclusive oversight, which they believe should shield it from state regulations, as stated in court documents. However, Woodbury pointed out that the question of whether federal law supersedes state law remains unresolved for the time being, though the courts have not shown a tendency in that direction.

Kalshi did not provide a comment on the situation when approached by TechCrunch. Wired was the first to report on the judge’s ruling. According to Reuters, Nevada had previously persuaded judges to impose bans on Kalshi’s competitors, such as Coinbase and Polymarket.

The Nevada litigation involving the prediction market is just one among a growing number of state cases nationally that argue that platforms such as Kalshi and Polymarket are illegal operations that evade state gambling regulations.

In contrast, current federal officials have taken on the role of advocates for the prediction industry. A prime example: Following Arizona’s decision to file criminal charges against Kalshi earlier this week, the CFTC’s chairman, Mike Selig, publicly criticized the decision, stating online: “The Arizona Attorney General today filed criminal charges against one of our registered exchanges pertaining to prediction markets. This is a jurisdictional conflict and utterly unsuitable as a criminal prosecution. The @CFTC is monitoring this closely and assessing its options.”

The increasingly adversarial stance of state officials combined with the leniency from the CFTC has essentially assured a regulatory conflict between states and the federal government regarding prediction markets and their future.

The CEO of Pinterest, Bill Ready, has urged governments to prohibit social media usage for individuals under the age of 16 in a recent op-ed featured in Time.

He is now among the few prominent CEOs advocating for such a measure, as countries around the globe start to enact or contemplate strategies to curtail children’s online engagement and determine how to authenticate ages. 

Ready expressed that today’s children are “undergoing the most extensive social experiment in history,” having been granted “unrestricted access to social media platforms.” Studies indicate the harm this unregulated access has caused, with present-day youth experiencing higher levels of depression, anxiety, and diminished focus. 

He remarked that social media companies provided “inadequate consideration of the ramifications” their platforms could have on minors and commended Australia’s initiative to ban social media for children, stating that “if tech firms do not put youth safety first, other nations should imitate Australia’s example.”

He highlighted the restrictions and regulations imposed on sectors like tobacco and alcohol, asserting that “such measures can enhance and occasionally preserve lives.” He also mentioned that Pinterest has thrived with Gen Z, even after restricting access to the site’s social functionalities for users below 16 years old.

“When we rationalize our inaction regarding the public’s best interests, tech CEOs mirror the tobacco executives of the 20th century who needed to be coerced and litigated into compliance,” he stated.

Besides Australia, countries such as Malaysia, Spain, and Indonesia have declared social media restrictions. This initiative is igniting discussion within the technology sector, but policymakers show no inclination to yield to the pressure. Legislators in France recently sanctioned a ban for users under 15, while Germany’s ruling faction affirmed its support for a variant of a social media prohibition as well. Various states in the U.S. are also exploring methods to limit minors’ social media access.

A recently released survey comprising over 500 science conference presentations over two years aimed to assess whether scientists possess a sense of humor, which is amusing in itself, if not the most effective use of time. The findings were somewhat predictable: two-thirds of humor attempts received either mild chuckles or complete silence, with only 9% resonating sufficiently to ignite laughter among most attendees. The most significant laughter, unsurprisingly, stemmed from technical glitches, such as malfunctioning slides and microphones cutting out. (Nothing unifies an audience faster than witnessing someone else’s mishap.)

Anyone who has endured a conference on any subject, anywhere, recognizes that scientists are not the only ones who can flop. Humor is challenging to execute in front of an audience that hasn’t been warmed up. Even SNL refers to its opening segment as a “cold open” — the audience hasn’t laughed at anything yet, making that initial laugh the most difficult to achieve.

Approximately 40% of the presentations completely steered clear of humor, which is a safe approach but likely leads to an even longer afternoon. More fascinating — according to science — is that it renders talks less memorable. “Despite the vast array of engaging content at conferences, it can be tough to remain focused. And by focused, I mean awake,” one physician-scientist mentioned to Nature, which also interviewed one of the study’s eight(!) co-authors.