To purchase this Bay Area residence, you will require Anthropic equity

To purchase this Bay Area residence, you will require Anthropic equity

An intriguing offer is on the table for a 13-acre estate located in Mill Valley, just to the north of San Francisco.

Storm Duncan, the homeowner and investment banker, has launched a LinkedIn profile for the residence, stating that he’s “interested in exchanging […] for Anthropic equity.”

–the San Francisco Standard notes that Duncan referred to this as a “diversification strategy,” explaining that he is “under-invested in AI related assets given AI’s future significance, and over-invested in real estate,” while a younger employee at Anthropic might find themselves “in a precisely opposite situation.”

Duncan is inviting interested parties to contact him via email to go over the details of the transaction, but he indicated that it would be a private deal that does not necessitate the buyer to liquidate their stock immediately. On LinkedIn, he also mentioned that the homebuyer would “keep 20% of the potential increase in value of the shares traded throughout the lockup duration.”

Duncan, who has described himself as a long-time resident of the Bay Area who relocated to Miami during the pandemic, acquired the property in 2019 for $4.75 million. He mentioned that it is currently inhabited by “a notable VC,” but he opted not to reveal the identity of the VC.

SpeakON’s dictation tool is a solid concept hindered by platform restrictions

SpeakON’s dictation tool is a solid concept hindered by platform restrictions

I frequently utilize dictation applications such as Wispr Flow, Willow, or Typeless to respond to texts and emails on my Mac and phone. However, for this, I must rely on my phone’s microphone or AirPods to dictate my messages, which often fail to capture my voice accurately.

That’s why I was curious about the proposal from Notta-owned SpeakON to try a dedicated dictation device. Sadly, I encountered mixed experiences with the product due to its design and platform constraints.

Nonetheless, I believe there is potential for dictation devices like this in the future.

SpeakON resembles a small pebble that attaches to the back of your iPhone via MagSafe, similar to Plaud’s AI meeting notetaker. It is very lightweight at 25 grams, so you won’t notice its presence, even if you carry it separately in your pocket.

Image Credits:Ivan Mehta

The device includes a companion app for iOS that functions as a keyboard, similar to other dictation applications. You may also use the app independently of the device if necessary.

To begin using it, simply press the button on the device to start dictation and release it when finished. The device features a single microphone and claims to capture audio from up to two feet away. It works with any app as long as the software keyboard is active. The speech is automatically filtered as the app eliminates filler words and can format the output as a list if required.

One benefit is that the device does not utilize the iPhone’s microphone, depending instead on its own. Other dictation apps require the iPhone microphone to be active based on user-defined session times. The SpeakON device helps you bypass that.

My concern with the SpeakON device is that, despite having dedicated microphones, it struggles to capture audio properly—unless the phone is about two feet away. Within that distance, the microphones frequently underperform due to background noise. I hope the next version will feature improved microphones.

I also wish that double-tapping the record button could bring the SpeakON keyboard to the front if I were using a text keyboard. Or if I could begin speaking without needing to switch keyboards, but those appear to be system-level limitations that might be difficult to resolve.

The app adjusts AI editing and tone based on your current app, but manual adjustments are also possible. At times, I found the edits to be forced and unnecessary. For example, when I asked, “Does this app work automatically?”, the app changed it to, “Does this application operate automatically?” In another case, it altered “complex” to “tricky,” and “Sure, no worries” became “There is no need to be concerned.” I don’t desire or require these types of edits. I eventually disabled its “attune” tone-changing feature for improved results.

I would have utilized this device significantly more if it had been compatible with Mac, allowing me to dictate in any application.

Image Credits:Ivan Mehta

In addition to transcription, users can hit the translate button to automatically convert speech into supported languages, including English, Japanese, Korean, Simplified Chinese, Traditional Chinese, Spanish, French, German, Italian, Portuguese, Russian, and Arabic.

SpeakON claims that the device can be used continuously for 10 hours and has a standby time of 20 days. However, in my experience, standby lasted only a few days instead of 20. By default, the device remains on all the time. I recommend adjusting that setting to make the device turn off after a few hours without use to conserve battery life.

The device can fully recharge from 0 to 100% in an hour. However, you may also plug it in for a few minutes to obtain enough power for several minutes of dictation.

The device is available for $129 with a plan allowing 5,000 words per week. Other apps like Wispr Flow usually provide 2,000 words per week on their free plan. There’s also a $12 monthly plan for unlimited words.

SpeakON has a first-mover advantage by introducing a dictation device. However, it must enhance platform compatibility and improve the software experience, as another company could easily procure components and create a different form factor to capture market share.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

'Saros' Highlights PS5's DualSense Capabilities

‘Saros’ Highlights PS5’s DualSense Capabilities

**Spoiler Alert for the Opening Scene of the Upcoming Game Saros:** The game kicks off by showcasing a sequence of words. These characters materialize on the screen one at a time, presenting certain world-building aspects to lay the groundwork for the narrative. The exact details are easy to miss because of the strong tactile experiences prompted by the controller. Each letter evokes a sharp haptic buzz, creating a clicky-clacky sensation. From the very beginning, Saros emphasizes its commitment to immersive storytelling through physical experiences.

Since the launch of the PlayStation 5, its DualSense controllers have enabled developers to harness haptic feedback, allowing the controller to vibrate in targeted ways. This innovation can replicate various on-screen sensations, such as typing letters, the sound of rain, or the powerful tremor when discharging a weapon or executing a melee strike. The adaptive triggers offer varying levels of resistance, clarifying the difference between a light pull and a full press.

Saros, scheduled for release on April 30, is crafted by Housemarque, a Finnish studio under Sony’s ownership. The studio previously introduced the widely praised PlayStation 5 title Returnal in 2021. As a launch title for the console, Returnal highlighted the newest advancements in Sony’s hardware, particularly the haptic and adaptive features of the DualSense controller. Gregory Louden, the creative director at Housemarque, managed the development of both titles and felt a distinct pressure to showcase the console’s potential.

“Releasing Returnal, there was a feeling of obligation since it was a launch title for PlayStation 5. We aimed to explore what the hardware could accomplish,” Louden revealed to WIRED. “We develop for the audience, but we also seek to advance the gaming medium and inspire others.”

Similar to Returnal, Housemarque’s new installment takes full advantage of the PlayStation 5’s DualSense controllers. It also incorporates 3D audio to enrich the game’s atmosphere. Although both titles operate on the same hardware, Louden is confident that the integration and experience are now more fluid.

“We pushed the graphics and hardware boundaries,” Louden articulated. “Our ambition was to provide something exceptional to players and fully leverage the potential of the DualSense.”

After several hours of gameplay, Saros emerges as a captivating experience. This dark sci-fi roguelike features intense confrontations where players fend off swarms of ominous aliens with dynamic, tactile action. The immersion is amplified by the synergy between what is displayed and the feedback from the controller. Clear mechanics like firing weapons or absorbing hits with a shield through haptic feedback are augmented by subtler techniques, such as during cinematic moments, where steady haptic pulses heighten the tension and emotions conveyed by the characters.

Anthropic developed a testing platform for agent-to-agent transactions

Anthropic developed a testing platform for agent-to-agent transactions

In a recent study, Anthropic established a classified marketplace where AI agents acted on behalf of both buyers and sellers, executing genuine transactions for actual goods and real currency.

The organization acknowledged that this trial — referred to as Project Deal — was merely “a pilot experiment with a self-selected participant pool” consisting of 69 Anthropic staff members who were allocated a budget of $100 (distributed through gift cards) to purchase items from their colleagues.

Despite this, Anthropic remarked that it was “impressed by the effectiveness of Project Deal,” with 186 transactions completed, amounting to over $4,000 in total value.

The company noted that it actually conducted four distinct marketplaces utilizing different models — one that was “real” (where every participant was represented by the company’s most advanced model, and with agreements genuinely honored post-experiment) and three others for research purposes. 

Evidently, when participants are represented by more sophisticated models, they achieve “objectively better outcomes,” according to Anthropic. However, users seemed unaware of the discrepancy, suggesting the possibility of “‘agent quality’ gaps” where “individuals at a disadvantage may not realize their inferiority.”

Moreover, the preliminary instructions provided to the agents did not seem to influence the likelihood of sales or the prices negotiated.

Maine’s governor rejects data center moratorium

Maine’s governor rejects data center moratorium

Maine’s Governor Janet Mills has rejected a bill that would have temporarily paused permits for new data centers.

Had it been enacted, L.D. 307 would have established the nation’s first statewide freeze on new data centers — effective until November 1, 2027. The legislation also proposed the formation of a 13-member council tasked with analyzing and recommending actions regarding data center development.

With mounting public disapproval of data centers, other states, such as New York, have contemplated similar temporary bans.

In correspondence to the state assembly, Mills — a Democrat currently campaigning for the U.S. Senate — stated that halting new data centers would be “suitable considering the effects of large data centers in other states on the environment and electricity costs” and that she “would have approved this bill” if it had provided an exception for a data center initiative in the Town of Jay.

That initiative, according to Mills, “has substantial local backing from its host community and surrounding area.”

Melanie Sachs, a Democratic state representative who introduced the legislation, remarked that Mills’ veto “poses considerable potential ramifications for all consumers, our electrical grid, our environment, and our collective energy future.”

Key Focus Applications of 2026: Focus Buddy, Forest, Focus Voyager

Key Focus Applications of 2026: Focus Buddy, Forest, Focus Voyager

When attempting to maintain concentration, distractions from your phone, web browser, or outside can be abundant. With attention spans diminishing in the TikTok age, there exists an entire category of applications aimed at helping you stay on task.

These applications operate in a similar manner, providing an easy method to monitor the time spent on an assignment and offering motivation to persist for the designated duration. Some features include the ability to block access to other applications.

To create this particular article without frequently switching between browser tabs and applications, I evaluated three leading focus tools. Here’s how they stack up.

Focus Friend

Focus Friend delivers a warm, inviting experience. Upon initial setup, you are given a personal bean to name. After going through the introductory screens, your bean begins to knit—and will stop if you pick up your phone.

If you keep your focus session intact, your bean crafts various knitted creations, which can be traded for home decor in its living area. Additionally, you can buy decorations, and a Pro subscription ($2/month) grants access to more creative knitting options.

You determine the duration of focus sessions and can play soothing music, block other applications, or keep the screen active while working or studying as your bean knits. The setup is charming and straightforward, requiring no account registration.

The effectiveness of Focus Friend relies on your connection to the bean and its projects. Although it essentially serves as a stopwatch, the app’s capability to block other applications is beneficial. The companionship aspect aids in improving my concentration.

Focus Friend for Android and iOS (free or from $2/month)

Forest

Forest helps you maintain focus by cultivating virtual trees. The longer you remain free from distractions, the more trees you collect, eventually forming your own forest on your phone. Forest has teamed up with the nonprofit Trees for the Future to plant millions of real trees globally.

OpenAI's CEO expresses regret to the Tumbler Ridge community

OpenAI’s CEO expresses regret to the Tumbler Ridge community

In correspondence addressed to the inhabitants of Tumbler Ridge, Canada, OpenAI CEO Sam Altman expressed his “deep regret” that his organization did not notify law enforcement regarding the suspect in a recent mass shooting incident.

Following the identification of 18-year-old Jesse Van Rootselaar as the alleged shooter who reportedly took the lives of eight individuals, the Wall Street Journal revealed that OpenAI had flagged and suspended Van Rootselaar’s ChatGPT account in June 2025 for discussing scenarios related to gun violence. The company’s team debated notifying authorities but ultimately chose not to, reaching out to Canadian officials only after the tragedy occurred.

OpenAI has since stated that it is enhancing its safety measures, such as implementing more adaptable criteria for determining when accounts should be reported to authorities, and creating direct communication channels with Canadian law enforcement.

In Altman’s letter, initially published in the local newspaper Tumbler RidgeLines, the CEO mentioned having conversations about the shooting with Tumbler Ridge Mayor Darryl Krakowka and British Columbia Premier David Eby, and they collectively agreed “a public apology was essential,” but “time was also needed to honor the community during their mourning.”

“I am profoundly sorry that we did not inform law enforcement regarding the account that was suspended in June,” Altman remarked. “While I understand words may never suffice, I feel an apology is essential to acknowledge the pain and irrevocable loss your community has endured.”

Altman further indicated that OpenAI will “persist in collaborating with all levels of government to help ensure that nothing like this occurs in the future.”

In a statement on X, Eby noted that Altman’s apology is “necessary, and yet markedly inadequate for the destruction caused to the families of Tumbler Ridge.”

Techcrunch event

San Francisco, CA
|
October 13-15, 2026

Canadian authorities have indicated they are contemplating new regulations regarding artificial intelligence but have yet to reach any conclusive decisions.

When you buy through links in our articles, we may receive a small commission. This doesn’t affect our editorial independence.

The climate tech IPO opportunity might finally be starting to open up.

The climate tech IPO opportunity might finally be starting to open up.

Startups focused on climate technology are highly reliant on capital, have extended timelines, and their innovations are frequently regarded as “novel.” Furthermore, a primary selling point is the mitigation of pollution — an externality that, at its best, is inadequately valued by the market. These characteristics are typically not favored by stock investors.

Nonetheless, the public markets seem to be becoming more receptive to climate tech startups — or at least a select few.

This week, nuclear venture X-energy debuted publicly, garnering $1 billion through an expanded share offering that seemingly provided substantial returns for its backers, including Amazon. Retail investors appear to be eager, with the stock surging 25% during its initial trading hour. Additionally, geothermal company Fervo announced it has applied for an initial public offering. The details of the Fervo IPO remain undisclosed, but private investors have appraised the firm at approximately $3 billion, as reported by PitchBook.

The decision to go public corresponds with what investors shared with TechCrunch late last year. After years of lukewarm interest in climate tech firms, they anticipated public markets to begin embracing energy-related startups. Almost every investor consulted on the matter indicated that startups with the highest potential for going public focus on nuclear fission or improved geothermal technology. Fervo, in particular, was referenced multiple times.

Credit data centers for this shift. The AI trend has transformed a growing demand for electricity into an appealing and marketable narrative. Companies that were already positioned to benefit from the increase found themselves aligned with a compelling story that matched their technological advancements. Luck certainly favors the prepared.

The IPOs are also likely to please investors, enabling them to return funds to their limited partners. The recent scarcity of IPOs has kept a portion of climate tech financing tied up, during a time when many funds are eager to begin cashing out.

However, it isn’t solely about cashing out.

Techcrunch event

San Francisco, CA
|
October 13-15, 2026

Fervo and X-energy have opted for the conventional path to the public markets, indicating there is confidence in a broad demographic of investors wishing to engage. If the goal were merely to unlock investor capital, the startups could have pursued the SPAC approach. (Several have.) Yet these two enterprises chose the more prolonged route.

Despite their accomplishments, a significant portion of climate tech may ultimately miss out on this IPO surge.

Firms not involved in energy markets will need to discover alternative methods to continue advancing — all while lacking access to the substantial resources provided by public markets. This divergence hints that the climate tech landscape is beginning to adopt a K-shaped trajectory, a notion Mark Cupta, managing director at Prelude Ventures, proposed when I spoke with him just over a week ago.

Companies remaining on the less affluent side of the IPO spectrum can still turn to private investors for support. However, a K-shaped trend seems to be emerging there as well.

According to Sightline Climate, venture capital and growth funds accumulated roughly $6.5 billion last year. This mirrors the amount in 2021, but due to the rise in the number of funds, each fund is now smaller. This could spell trouble for founders, as funds have fewer resources available. On the positive side, increased competition might lead to improved fundraising outcomes.

Simultaneously, larger funds continue to expand. Infrastructure dominated climate tech investment last year, with 42 funds accounting for 75% of all capital in the sector, according to Sightline Climate. This success is expected to transfer to the startup sector if the company possesses a mature technology ready for large-scale development.

Sightline noted that many new infrastructure funds are concentrating on renewables, grid solutions, and energy storage. In essence, the K-shaped phenomenon is unlikely to dissipate anytime soon.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

Reasons for Cohere's merger with Aleph Alpha

Reasons for Cohere’s merger with Aleph Alpha

The Canadian AI firm Cohere is acquiring Germany’s Aleph Alpha with backing from Schwarz Group, the parent corporation of Lidl supermarkets. With governmental support, these companies aim to provide a sovereign alternative for businesses in an AI environment largely influenced by American firms.

Both Aleph Alpha and Cohere are known for their development of large language models, having gained recognition in their local markets, yet they still trail behind global leaders like OpenAI. Despite some common ground, this collaboration is not between equals. Cohere, which was last valued at $6.8 billion, will head the new organization that will merge Aleph Alpha, pending regulatory and shareholder approvals.

Schwarz Group, a primary investor in Aleph Alpha, has already expressed full support for the acquisition. The retail powerhouse will become a strategic partner of the new organization with an investment of €500 million in structured financing (around $600 million) — and anticipates leveraging STACKIT, the sovereign cloud solution from its IT arm, Schwarz Digits.

In conjunction with its financial contribution, Schwarz Group is also serving as Cohere’s lead investor in the Series E funding round, having set the valuation at approximately $20 billion. According to the German news outlet Handelsblatt, the term sheet puts this valuation firmly in place.

This would represent a notable advancement that can’t solely be justified by combined revenue numbers. While Cohere recorded $240 million in annual recurring revenue in 2025, Aleph Alpha had previously seen minimal revenue and considerable losses. Nonetheless, investors are optimistic that this partnership will enhance their prospects.

They may share this belief with others. Elon Musk’s AI venture xAI has reportedly contemplated a three-way partnership with France’s Mistral AI and Cursor, which SpaceX has recently shown interest in acquiring. However, it remains uncertain whether the French company would risk jeopardizing its position as an alternative to U.S. technology, which has been beneficial for its revenue streams.

Cohere is similarly looking to capitalize on enterprises seeking alternatives to AI service providers that may not fulfill their standards for privacy and autonomy. The new establishment plans to focus on highly-regulated sectors such as defense, energy, finance, healthcare, manufacturing, telecommunications, and the public sector.

Techcrunch event

San Francisco, CA
|
October 13-15, 2026

Aleph Alpha has also created specialized language models oriented toward businesses and public entities in Europe, like the PhariaAI suite. A later strategic shift and the exit of co-founder and CEO Jonas Andrulis has left its strategy and leadership somewhat ambiguous, yet its team of 250 skilled professionals could still enhance Cohere’s capabilities.

“Their emphasis on small language models, European languages, and tokenizers is very complementary to our approach, which focuses more broadly on large language models,” stated Cohere CEO Aidan Gomez during a press conference announcing the partnership on Friday.

The event’s guest list was also noteworthy. Instead of Aleph Alpha’s co-CEOs, co-founder Samuel Weinbach took the stage with Gomez alongside Schwarz Group’s chief digital officer Rolf Schumann. The gathering also included German digital minister Karsten Wildberger and his Canadian counterpart Evan Solomon.

As tensions rise with the United States, Canada is increasingly eager to form bilateral partnerships with various nations, including Germany. Sharing concerns over privacy and security, the two nations recently established a Sovereign Technology Alliance aimed at “enhancing sovereign AI capabilities and diminishing strategic technological dependencies.”

The lingering question is whether European organizations will perceive a Canadian initiative as sufficiently sovereign, or if they will trust that the alliance will remain transatlantic in the long term. According to Gomez, “Cohere will become a Canadian-German company.” Nevertheless, ownership may soon become less defined if a public offering is on the horizon.

When you make purchases through links in our articles, we may earn a small commission. This does not influence our editorial integrity.

Why Tokyo stands as the premier tech hub of 2026

Why Tokyo stands as the premier tech hub of 2026

Every significant tech event revolves around themes. Many are ambiguous enough to convey everything and nothing simultaneously. SusHi Tech Tokyo 2026 is taking a different approach — four precisely defined technology sectors, each supported by live demonstrations, dedicated exhibit areas, and sessions featuring individuals who are actively developing and financing these technologies worldwide.

TechCrunch is collaborating with SusHi Tech Tokyo 2026 as an official media partner, and our Startup Battlefield team will be on-site choosing one exceptional semifinalist from the SusHi Tech Challenge to move forward to the TechCrunch Disrupt Startup Battlefield 200 — one of the most esteemed launchpads in the tech industry. Here’s what you can expect on the floor.

AI — beyond the hype, into the infrastructure

Sessions featuring Howard Wright (Nvidia), Rob Chu (AWS), and Eric Benhamou (Benhamou Global Ventures) will address the actual deployment of AI on a large scale and identify the genuine risks involved. On the floor, university startups focused on AI pitch alongside international players, with the AI Film Festival Japan, a partner event at Tokyo Innovation Base in Yurakucho, examining how artificial intelligence is actively transforming culture.

Robotics — physical AI has arrived

The robots at SusHi Tech are not behind enclosed spaces — they are present on the floor and engaging. On stage, Nissan, Isuzu, and Applied Intuition’s Qasar Younis explore how software-defined vehicles are revolutionizing transportation. Physical AI is not a future trend; it’s here in Tokyo on April 27.

Resilience — the cities that survive what’s coming

Eva Chen (Trend Micro) and NEC’s Noboru Nakatani address cyber defense, while leading climate tech VCs from Breakthrough Energy and Cleantech Group investigate global investment trends. A VR disaster simulator and tours of Tokyo’s underground flood-control systems bring the stakes to life.

Entertainment — Japan’s cultural engine meets AI

Sessions with the CEOs of Production I.G, MAPPA, and CoMix Wave Films discuss the essentials for Tokyo to evolve into the Hollywood of animation. On the exhibition floor, startups leverage AI to translate manga globally, create music from text prompts, and animate Japanese IP for worldwide distribution.

Can’t make it to Tokyo? You can still be there

Missing SusHi Tech Tokyo doesn’t mean you have to miss out. Remote participants receive more than just a livestream — on-site staff will navigate the floor on your behalf, equipped with a device displaying your likeness so you can engage with attendees and exhibitors in real-time, face-to-face. It’s the closest experience to being there.

Techcrunch event

San Francisco, CA
|
October 13-15, 2026

Note: Please bear in mind that some sessions may not be available for viewing.

Apply for remote participation with assistance from on-site staff here.

Can’t manage that either? Ticket holders have the option to stream sessions online and access the programming from anywhere. Check out the complete session list here.

Alongside the startup event, the Tokyo Metropolitan Government also facilitates a gathering of leaders from 55 cities across five continents. They will discuss “A New Urban Future Built on Climate and Disaster Resilience.” The summit for city leaders is part of G-NETS (Global City Network for Sustainability), organized by the Tokyo Metropolitan Government since 2022 as a forum to address common challenges with a focus on resilience to urban climate disasters and citizen well-being. The summit will be viewable by general audiences on YouTube live and after the event.

G-NETS official website

G-NETS YouTube Channel 

SusHi Tech Tokyo 2026 will take place from April 27–29 at Tokyo Big Sight. Business days are April 27–28; public day (free admission) is April 29. Click here to register.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.