AXPONA is over. Passover is done. The Stanley Cup Playoffs are underway, which is always a beautiful time of year unless you’re a Toronto Maple Leafs fan, in which case it’s just another annual reminder that hope is a renewable resource and results are not.

So I did what any rational person does. I got in the car and tried to drive.

Three hours later I was still in New Jersey, locked in traffic like it was a civic duty, leaning on Qobuz and the NHL Network to keep me from turning the steering wheel into modern art. The stereo in my Toyota SUV did not help. It is not just bad. It is hostile. Flat, lifeless, and oddly aggressive in how it strips the soul out of anything you feed it. I have heard better sound quality from hold music and from relatives calling to critique my life choices. At least they bring some midrange.

Which is why the arrival of the 2026 Mazda CX-5 Premium cannot come soon enough. The Bose system alone is reason to celebrate. Clarity. Separation. Actual bass that does not feel like a rumor. A car where music sounds like music and not a compressed apology for itself.

Because music matters. Especially when you are stuck and have no escape.

Country is also where things fall apart. What passes for country today feels like it was assembled by committee in a conference room with a whiteboard and no sense of shame. I cannot do it. I will not do it. There are exceptions. Orville Peck gets a pass. Dolly Parton from an earlier era still hits. Tom Russell knows how to tell a story without sounding like he is selling me a truck.

And while we are clearing the table, let’s deal with the sacred cows. Eagles. I do not get it. Never have. Polished to the point of anesthesia. Give me Led Zeppelin, The Cure, or Guns N’ Roses any day.

And then there is Glen Campbell, who was not just good, he was essential. You can roll your eyes at “Rhinestone Cowboy,” fine, but “Wichita Lineman” is something else entirely. “I need you more than want you” is not just a lyric, it is a confession. Every word lands. It does not matter if you have never set foot in Kansas, although I have. The song pulls you there anyway. It makes you feel distance, longing, and the quiet weight of holding on to something that might already be gone, but hopefully isn’t.

That is what music is supposed to do. It reaches across time and geography and lands somewhere personal. Somewhere you do not always expect. In my case, it took me halfway across the world for a few minutes. Tokyo. A Japanese egg salad sandwich from 7-Eleven. Blonde hair. A mind that can navigate the dark side of the forest moon and come back with answers. Present without being present.

And it is also why, when you start looking at where some of our favorite stories really came from, you have to be willing to follow that thread wherever it leads.

The Kurosawa Blueprint That Built Star Wars

A long time ago, in a galaxy not so far away, Japan, a director named Akira Kurosawa made a film in 1958 that quietly rewrote the playbook. The Hidden Fortress did not need a marketing machine or a line of action figures. It just told a story with precision, perspective, and a structure that would echo a lot louder years later.

Fast forward to 1977. I am seven years old, sitting in a theater in Toronto, watching Star Wars with my friend Andrew Temes, completely unaware that my cinematic worldview was about to be rearranged. Like a lot of people, I did not just like it, I went all in. Thousands of dollars on memorabilia over the years. Stood in line, in the rain, to be the first one through the door in Canada for Return of the Jedi in 1983. Fast forward again and there I was in my late twenties, back on line for The Phantom Menace, which felt less like a return and more like a warning shot. Outside of Andor, Star Wars: The Clone Wars, Rogue One, and maybe Solo: A Star Wars Story, it has been a long stretch of Bantha-level disappointment since.

Then I found Kurosawa.

Throne of Blood. Seven Samurai. Yojimbo. High and Low. The Bad Sleep Well. One by one, they chipped away at that original belief. I own them all now, in more formats than I care to admit. But it was not until my twenties, sitting down with The Hidden Fortress, that the illusion finally cracked.

Because this is not subtle.

The two bickering peasants framing the story? They are not just an influence — they are the blueprint for C3PO and R2-D2. The hidden princess, the reluctant general, the journey through enemy territory, the tonal shifts between danger and dark humor, it is all there. And visually, it is even harder to ignore. The samurai swords, katanas, are not just cousins to lightsabers, they are the DNA. The code of the samurai maps cleanly onto the Jedi. The warlord’s soldiers, with their helmets and rigid formations, look a little too familiar. Strip away the desert, add black armor, and you are staring at the Empire before it learned how to breathe through a mask.

Look, every director borrows. That is part of the job. Influence is how the medium evolves. But this goes further than influence. This is Imperial level theft.

It does not take anything away from what Star Wars became. It changed cinema. It changed my life. But it did not come from nowhere. And once you see where it came from, you cannot unsee it.

And here is where it gets uncomfortable. Han Solo suddenly feels a lot less original and a lot more like every reluctant hero Toshiro Mifune ever played. From the early postwar Kurosawa films to the sprawling samurai epics to even Red Beard, that swagger, that resistance, that eventual turn toward doing the right thing, it is all there.

Which raises the obvious question. What exactly did George Lucas bring to the table beyond scale and spectacle?

And while we are asking questions, let’s talk about endings. Kurosawa knew how to stage a death. Watch Throne of Blood. Arrows flying, body breaking down in real time, a human pin cushion before the term even existed. It is brutal, precise, unforgettable.

Lucas? Not exactly his strength.

Grief, Guilt, and the Monster We Deserved

Sitting down for Godzilla Minus One, I expected another traditional monster movie. I could not have been more wrong.

Set in the final days of World War II and the fragile years that follow, the film strips the franchise down to something far more uncomfortable than spectacle. This isn’t about cities getting flattened for fun. It’s about what’s left standing after everything that matters is already gone.

Kōichi Shikishima isn’t a hero. He’s a failed kamikaze pilot who makes a choice to live and then has to carry the weight of it. When he lands his Mitsubishi Zero on Odo Island claiming mechanical failure, the lie hangs in the air. The mechanic, Sōsaku Tachibana, sees right through him. And when Godzilla attacks that night, Shikishima freezes. He doesn’t fire. He survives. Everyone else doesn’t. That’s the wound the film never lets heal.

When he returns home, Tokyo is gone. His parents are gone. What he builds instead, a fragile, makeshift family with Noriko and the orphaned Akiko, feels less like a new beginning and more like borrowed time. He takes work on a minesweeper, cleaning up the literal leftovers of the war, while internally he’s still stuck in it.

And then Godzilla comes back.

Not as a metaphor, not as a mascot, but as something far worse. Mutated and supercharged by American nuclear testing at Bikini Atoll, this version of Godzilla is a consequence. A continuation. The war didn’t end. It just changed shape. The U.S. steps back because of geopolitical tension. The Japanese government stays quiet to avoid panic. So the people left behind — the same ones already scraping together a life, are the ones who have to deal with it.

That’s the film.

Not the destruction, though there’s plenty of it. The Ginza sequence is devastating, not because of scale, but because of who gets caught in it. Noriko’s death isn’t a plot device. It’s the final break. Shikishima loses the one thing tethering him to something resembling a future, and what’s left is grief with a target.

The plan to stop Godzilla with Freon tanks, pressure, and improvisation feels almost secondary. It’s clever, grounded, and very Japanese in its resourcefulness, but the real story is whether Shikishima can find a reason to live that isn’t tied to dying for something.

That’s why this film works on a level that most Hollywood films do not. Kurosawa would have complained about the lighting in some scenes, but the tone and overall theme would have felt very familiar.

Made for a reported $10 to $15 million, it does more with restraint than the recent Hollywood entries have done with budgets ten times that size. The monster is used sparingly, and every appearance matters. There’s no filler. No winking at the audience. Just tension, consequence, and a very clear understanding of what Godzilla was always supposed to represent.

Seventy years into the franchise, this is the one that finally circles back to the source. Postwar trauma. Nuclear anxiety. Survivor’s guilt. The politics of rebuilding when nobody wants to admit how broken things still are.

I’ve seen almost all of them. Own more than I probably should.

This is the one that broke me in the theater after a year of personal upheaval and more than a few moments where doubt, loss, and betrayal had the upper hand.

Not handing it the Oscar was all you need to know about Hollywood.

The Last Honest Meal in a Crooked Economy

You know what’s actually life-affirming in the way Godzilla Minus One sneaks up on you? Not the destruction, but the rebuild. The small, human stuff that still works when everything else feels broken.

A good hot dog.

And here’s the part nobody wants to say out loud: hot dogs are coming back because everything else got stupid expensive. Pizza pushing $30 a pie isn’t a luxury anymore — it’s a bad habit with a pocket full of receipts. Meanwhile, the hot dog never left its lane. Still relatively affordable. Still fast. Still satisfying in a way that doesn’t require a loan from your kid with 300,000 Instagram followers or a wine pairing.

AXPONA had rooms that cost more than most people’s homes, but step outside that world and reality looks a lot more like a paper tray and a line out the door. You don’t “elevate” a Chicago hot dog. You respect it. A proper char dog like the ones at Portillo’s, comes loaded the way it’s supposed to: tomato wedge, pickle spear, peppers, celery salt, mustard. No ketchup unless you’re looking to start a fight.

Back home in Jersey, that same no-nonsense energy shows up in different forms. Hiram’s Roadstand is the first stop in Fort Lee, which is better known these days for its Korean BBQ. Some of the greatest jazz musicians of the last century walked out of Van Gelder’s studio, sessions done, ears ringing, and headed straight down Route 9W to Hiram’s for something healthier than a cigarette and bottle of gin.

Flat-top dogs, nothing fancy, no attempt to impress anyone who doesn’t already get it. Hiram’s doesn’t care about your macros or your cleanse. It exists because it works.

But somewhere along the way, curiosity or maybe poor judgment, pulls you deeper into North Jersey’s hot dog hierarchy. That’s how you end up at Rutt’s Hut.

Rutt’s isn’t just a stand. It’s a Jersey institution. The deep-fried “ripper” doesn’t try to charm you. The casing splits, the edges land at that exact point between crisp and snap, and suddenly you understand why people make the drive. For years, I kept my distance. Fried? Messy? Felt like a late-night decision best avoided. Turns out, I was late to the party.

The first bite resets expectations. Not refined. Not chasing any TikTok trends. The kind of food that doesn’t need explaining because it’s been right for decades.

And then the supporting cast shows up. Onion rings done the way they should be — no gimmicks, just the right amount of crisp. Add their homemade relish and mustard, and now you’re in deeper than you planned. It’s the kind of combination that makes everything else on the table feel optional. Better than Coke, honestly. And not the kind you drink sitting on the hood of your car in the parking lot of a 7-Eleven in Long Branch.

Now I understand why the wise guys made the detour to Clifton on their way to the Bing.

The System Says Sit Down. Mr. McMurphy Says Not Today

Driving home from Rutt’s Hutt takes patience these days because New Jersey is having a moment. Call it Hollywood East if you want. I live less than two miles from the new Netflix Studios Fort Monmouth, a $900 million project that’s already changing the rhythm of the Shore. Ten thousand people are expected to land here over the next couple of years, and some days it feels like half of Brooklyn beat them to it. The quiet streets by the ocean are not so quiet anymore. Sure, the house is worth a lot more now. So are the property taxes.

And when Hollywood comes knocking, waving $25,000 for two months of rent for some B-list actor you’ve never heard of, you don’t overthink it. You smile, take the check, and hope he doesn’t spiral into some performative meltdown after noticing the thirteen mezuzahs and the IDF Six Day War flag sitting behind glass in the office.

Part of that drive home from North Jersey takes me past a hospital where I spent time as a patient. Not visiting. Staying. Which is why One Flew Over the Cuckoo’s Nest never felt like fiction to me. It felt familiar. The nurse who made my life a misery. The routines. The slow grind of being told who you are supposed to be. I’ve talked about it publicly. No secrets left to protect.

At 56, I’m not interested in pretending that those chapters didn’t happen. They did. And it taught me something useful. You either let that kind of experience flatten you, or you learn how to direct it. I chose the second option. The energy that used to work against me now has a job. As Emperor Palpatine would say, it gives me “focus” and a very different perspective.

That doesn’t always sit well with people. I’m not built to fall in line or nod along when something doesn’t pass the smell test. I’m not interested in selling the idea that six figure systems are normal, or that spending more on cables than a car is a rational decision for most people.

I’m not here to play Westeros court politics or nod along while the room pretends everything makes sense. I’m closer to Han Solo than anything polished and obedient — shoot first, deal with the fallout later. And there’s a little Arya Stark in there too. I keep a list.

Driving past that hospital, thinking about where I’ve been and where I just was—rooms at AXPONA stacked with mid six figure systems, cables dressed like they have their own security detail — it starts making sense.

Maybe we’ve got the labels wrong.

Because if “normal” now means telling people this is the entry point, that this is what it takes to belong, then maybe the problem isn’t the audience.

We’ve spent too much time talking at people like they showed up without the right jacket and pair of white leather sneakers, pricing them out before they’ve heard a single note, and wrapping it all in language that sounds more like a late night pitch than anything connected to music. That’s not passion. That’s insulation.

And here’s the part that should make people a little uncomfortable. Not everyone is willing to take the little white pill with the cup of water anymore and nod along.

Some of us remember what this is supposed to be. Sanity isn’t just a state of being. It has an address.

And more people in this industry need to start finding it.

Related Reading:

Are you a Verizon customer looking to get more value out of your mobile or home internet service? Well, keep reading, because I’ve got some news you’ll be interested in. Verizon offers a collection of free and discounted streaming services with many of its popular plans. These perks offer substantial discounts on streaming memberships, giving you plenty of ways to save money while still getting the entertainment you love.

Services such as Netflix and Disney Plus already deliver content at a competitive price. These Verizon perks make those savings even better.

Note that the bargains listed below are available to those with eligible mobile phone and home internet plans, unless otherwise specified. Scroll on for a list of the best free and discounted streaming perks for you.

Read more: AT&T vs. Verizon: Which of These 2 Major Carriers Is Best?

Netflix and HBO Max bundle

Whether you are a Verizon myPlan mobile customer or a myHome internet subscriber, you can choose this streaming service perk. If you have Fios Home Internet, 5G Home Internet or an eligible mobile plan, sign up for Netflix and HBO Max (basic with ads), bundled together at $10 per month for a total monthly savings of $9. (Netflix with ads is $8 per month, and ad-based HBO Max costs $11 per month).

Note: Beginning May 6, the cost of the Netflix and HBO Max bundle through Verizon will increase to $13 per month (saving you $7 per month).

If you tack on this bundle, you can also upgrade to an ad-free HBO Max plan and still get a Verizon discount. Pay an additional $7.50 per month for the Standard ad-free plan (saving you $3.50 total), or get HBO Max Premium (which comes with 4K and 100 downloads) for an extra $12 per month. These savings will automatically be applied for existing Verizon customers.

Disney Plus Premium

If you subscribe to a mobile plan with Verizon and want to stream all the Disney, Pixar, Marvel and Star Wars content your little heart desires, you will be able to access Disney Plus Premium — ad-free tier — free for six months. Verizon customers can take advantage of this deal by enrolling by midnight on May 31. 

Disney Plus, Hulu and ESPN

If you want more than just Disney Plus, Verizon offers the Disney Plus, Hulu and ESPN Select (With Ads) bundle for myPlan and myHome customers for $10 per month (typically $20 per month). The previously offered Premium ad-free version of this bundle perk is no longer offered by Verizon.

YouTube Premium

You can subscribe to YouTube Premium for $10 per month ($4 discount) if you have a Verizon Unlimited Ultimate, Ultimate Plus or Ultimate phone plan. Internet customers are eligible with 5G Home, LTE Home, Verizon Home Internet Lite and select Fios Home Internet plans.

Note: The YouTube Premium perk increases to $13 per month, beginning on May 13. Based on YouTube Premium’s recent price hike, you’ll still save $4 a month with this perk.

YouTube TV

If you subscribe to any Verizon home internet plan, you qualify for YouTube TV at $63 per month, after a $20 discount. To get this deal, you must enroll for a YouTube TV subscription from the MyVerizon Account View, download the app to your device and link your account to the streamer. You can create up to six user accounts. The discount lasts for six months, before it reverts to the normal $83 monthly cost. You must be a new YouTube TV subscriber to take advantage of the deal.

FOX One

You can get the new FOX One app for $15 per month with an eligible mobile or home internet plan. The FOX One app offers live and on-demand access to networks like FOX News, FOX Business and FOX Weather, along with sports programming from FS1, FS2 and the Big Ten Network. This is a $5 monthly discount from the regular $20 monthly cost. To take advantage of the perk, you must enroll through Verizon in an Unlimited Ultimate, Unlimited Plus or Unlimited Welcome phone plan or one of six internet plans, including 5G Home, 5G Home Ultimate, LTE Home or Home Internet Lite.

ESPN Unlimited

If you need your NFL fix and missed out on the Sunday Ticket on Us promo, Verizon now has another way to watch the sports content you love. Subscribe to a Verizon Fios TV package that includes the ESPN Network, and you will be eligible to access ESPN Unlimited at no additional cost.

Apple One

With an eligible Verizon mobile phone or home internet plan, you can take advantage of this Apple One perk. The deal offers the full Apple streaming package — which includes Apple TV, Apple Music, Apple Arcade and iCloud Plus — for $15 a month ($5 discount) for individuals and $20 a month ($6 discount) for families.

Meta is ‘clear that Instagram and Facebook are intended for people aged 13 and older’, said the company.

The EU has preliminarily found that Instagram and Facebook are in breach of the Digital Services Act (DSA) for failing to “diligently” identify and mitigate risks that children under 13 face when using these platforms.

The findings are in relation to an investigation the EU launched against Meta’s popular social media platforms in mid 2024 over concerns that Instagram and Facebook use algorithms that stimulate “addictive behaviour” in children.

Users need to be at least 13 years old to use Instagram and Facebook. However, the Commission found that the company’s own restrictions against underage usage don’t work.

It said that minors under 13 can enter false birth dates with no effective controls in place to check its validity. The measures Meta has put in also do not promptly identify under-13 users to remove their access, the EU added.

Meanwhile, the tools Meta offers to report underage users is “difficult to use and not effective”, the Commission said in its statement.

Meta also does not follow up on these reports, which allows underage users to continue using the service without any checks, the European authority found.

Moreover, the social media giant’s lack of enforcement “builds on an incomplete and arbitrary risk assessment” the EU said, “which inadequately identifies” the risk underage users face when accessing Instagram and Facebook.

Meta’s assessment contradicts “large bodies of evidence” from all over the EU which finds that roughly 10 to 12pc of children under 13 are accessing Instagram or Facebook, the authority said, while also disregarding “readily available scientific evidence” which indicates younger children are more vulnerable to potential harms caused by these services.

Meanwhile, last October, the EU, in a different preliminary ruling, found that Meta does not provide Instagram and Facebook users with simple mechanisms to notify illegal content or challenge content moderation decisions.

Meta disagrees with today’s findings. In a statement to SiliconRepublic.com, a spokesperson for the company said Meta is “clear that Instagram and Facebook are intended for people aged 13 and older”, adding that they have “measures in place to detect and remove accounts from anyone under that age”.

“Understanding age is an industry-wide challenge, which requires an industry-wide solution, and we will continue to engage constructively with the European Commission on this important issue.

“We continue to invest in technologies to find and remove underage users and will have more to share next week about additional measures rolling out soon,” the spokesperson added.

Today’s (29 April) results are based on an in-depth investigation by the EU that included an analysis of Instagram’s and Facebook’s risk assessment reports, internal data and documents, as well as the platforms’ replies to requests for information, the Commission said in a statement.

These, however, aren’t the Commission’s final views on the matter. If they are confirmed in its ultimate findings, the Commission could fine Meta as much as 6pc of its total worldwide annual turnover. Meta made more than $200bn in revenue in 2025.

“Meta’s own general conditions indicate their services are not intended for minors under 13. Yet, our preliminary findings show that Instagram and Facebook are doing very little to prevent children below this age from accessing their services,” said the EU’s executive vice-president for tech sovereignty, security and democracy Henna Virkkunen.

“The DSA requires platforms to enforce their own rules – terms and conditions should not be mere written statements, but rather the basis for concrete action to protect users, including children.”

The Commission wants Instagram and Facebook to change their risk assessment methodology to properly evaluate which risks arise on its platforms in the EU and how they manifest. It also wants Meta’s social media platforms to strengthen their measures to prevent, detect and remove minors under the age of 13 from their service.

According to the DSA guidelines, age estimation, which includes age verification, is seen as the appropriate measure to ensure the safety of minors. In order to be effective, all age-assurance technologies are required to be “accurate, reliable, robust, non-intrusive, and non-discriminatory”, the EU said.

The Commission, meanwhile, is continuing on with its investigations into Meta’s other potential breaches in relation to this investigation, including the assessment and mitigation of risks arising from the design of Facebook’s and Instagram’s online interfaces, which, it said, could be leading to “addictive behaviour”.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

Perplexity’s Comet browser is starting to make a lot more sense on the iPad. After bringing Comet to iOS users last month, the company is now adding proper iPadOS support, including multiple windows and Split View. The update is available now on the App Store, and it fixes one of the biggest gaps from the browser’s tablet launch.

Is Comet now practical enough for everyday iPad use?

Comet’s main draw is its built-in AI assistant, which lets users search, ask questions, summarize pages, and manage some web tasks inside the browser. The feature was useful for quick lookups on mobile, but the iPad version needed proper multitasking tools to make better use of the larger screen.

Split View now lets users keep Comet open beside another app, such as Notes, Mail, Pages, Slack, or a PDF reader. Multiple windows also make the browser more practical for research-heavy tasks, where a user may want one window for browsing and another for follow-up questions, summaries, or comparing information.

The update should make Comet easier to use for students, writers, researchers, and anyone who uses the iPad as a laptop replacement. It also gives users a stronger reason to try Comet as their main browser instead of opening it only for AI-assisted searches.

Can this help more users switch to Comet?

Perplexity made Comet free to use late last year, which likely helped more people try the browser as an alternative to Chrome or Safari. Since then, the company has continued improving the browser and bringing it to more platforms. We previously tested Comet as a Chrome replacement and found that its natural-language browsing approach changes how users move through the web.

For iPad users, this is a practical upgrade. Comet now works better with the apps people already use, which could make switching to Perplexity’s AI browser a lot easier.

Apple will be adding a slate of new Apple Intelligence features to the Photos app in the next major operating system update, including extend, enhance, and reframing options.

Apparently, Apple is developing a new suite of tools powered by Apple Intelligence. The tools, which will be made available on iOS 27, iPadOS 27, and macOS 27, are slated for release this fall, according to Bloomberg.

Currently, Apple only offers a single AI-powered feature: Clean Up. The feature has debuted to somewhat mixed reviews, with some users complaining that it often creates artifacts or fills areas with inaccurate detail.

The new features will include the usual slate of AI-powered photo editing features, like extending, reframing, enhancing, and contextual editing. For instance, enhance will automatically improve color, lighting, and image quality.

An AI-based extend feature would allow users to generate additional content beyond the frame. To do so, users will just need to extend the edges of an image with their fingers.

Reframe is the one feature that is notably different from the others. Designed to target spatial photos, users will be able to shift perspective after a shot is taken.

According to Bloomberg, the processing will be done on-device. That doesn’t come as much of a surprise, considering Apple has reiterated that it wants to avoid sending users’ data off-device whenever possible.

These features have been available for quite some time in other apps and on other platforms. The ability to extend photos, for instance, has been available in Photoshop for nearly three years.

Apple was always going to bring these features to market; it has to keep pace with competitors. However, Apple has struggled to release certain Apple Intelligence features in a timely fashion, so this timeframe may be tenuous at best.

Founded by a former Atomico talent adviser, Dex charges employers agency fees only on successful hires, and has gone from zero to $1.8 million ARR in under six months of charging.

Paddy Lambros spent two and a half years advising roughly 100 European startups on hiring at Atomico, one of Europe’s largest venture capital firms. The recurring lesson, was simple: nearly every serious problem at an early-stage company was downstream of a hiring decision. The wrong person in the role, or the right role left unfilled too long, and otherwise promising companies would stall.

That observation became Dex, an AI talent agent for technical hiring that Lambros founded in early 2025. The London-headquartered startup has raised a $5.3 million seed round led by Notion Capital, with participation from a16z Speedrun, Concept Ventures, and angel investors from OpenAI and other firms, the company announced on Monday.

The new funding brings total capital raised to $8.4 million, following a $3.1 million pre-seed last year.

What Dex does?

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

Dex has deliberately narrowed its focus to a single slice of the technical talent market: AI researchers, software developers, and machine learning and quantitative engineers.

More than 15,000 engineers have signed up to the platform, and over 50 technology companies, including Lovable, ElevenLabs, Synthesia, Granola, and Fyxer, are paying for the service.

Since beginning to charge in late 2025, Dex has grown from zero to approximately $1.8 million in annualised recurring revenue. Lambros described profitability by year-end as “conceivable.”

The product works in two stages. A candidate begins with a conversation, either voice or text, with Dex’s AI agent, which asks open-ended questions about their experience, motivations, and ambitions.

The agent, built on a combination of models from Google, Anthropic, and OpenAI, then surfaces relevant roles from a curated set of openings, helps the candidate research companies, benchmarks compensation, and prepares them for interviews.

On the matching side, Dex uses what Lambros describes as ‘old-school machine learning’, a proprietary engine built from the richer, more detailed profiles the conversational AI assembles, to surface candidates to employers. When both sides indicate mutual interest, Dex introduces the candidate directly to the hiring manager.

The business model is deliberately agency-shaped rather than SaaS-shaped. Dex does not sell software to recruiters or integrate with applicant tracking systems. It charges employers a success fee of 20–30% of a hired candidate’s first-year salary, the same fee structure as a traditional executive search firm. “We only earn money if we do a good job,” Lambros said.

The fee model is both a strategic choice and a pitch to employers who have grown wary of software subscriptions that promise to improve hiring without accountability for outcomes.

Lambros’ argument is that AI capable of sustained, contextually-rich back-and-forth conversation has, for the first time, made the core agency function genuinely automatable, and potentially better than the human version at scale.

A human recruiter knows a few hundred candidates; an AI agent can hold detailed conversations with hundreds of thousands. A human recruiter works with a handful of clients at once; an AI agent can serve thousands simultaneously.

The depth of data that candidates share privately with a conversational agent is also, Lambros argues, meaningfully richer than what they would put on a public LinkedIn profile or CV.

That data asymmetry is Dex’s core product thesis: the quality of the match is only as good as the quality of the candidate model, and most existing tools, LinkedIn’s recruiter products, cold-outreach scrapers, job boards, are working with shallow, public information.

The seed funding will go towards opening offices in New York and San Francisco later in 2026, taking the company into the most competitive market for AI engineering talent.

Lambros’ background, talent at Improbable when it grew from 50 to 650 people, people and talent at construction-tech startup Sensat, and then advisory work across Atomico’s portfolio, gives him credibility on both sides of the hire.

If a London-founded, vertically-focused AI talent agent can hold its positioning against better-capitalised generalists as both scale into the US market is the question the next twelve months will answer.

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has ordered federal agencies to secure their Windows systems against a vulnerability exploited in zero-day attacks.

Tracked as CVE-2026-32202, this security flaw was reported by cybersecurity firm Akamai, which described it as a zero-click vulnerability left behind after Microsoft incompletely patched a remote code execution flaw (CVE-2026-21510) in February.

As CERT-UA revealed, the Russian APT28 (aka UAC-0001 and Fancy Bear) cyberespionage group exploited CVE-2026-21510 in attacks against Ukraine and EU countries in December 2025 as part of an exploit chain that also targeted a LNK file flaw (CVE-2026-21513).

“Microsoft fixed the initial RCE (CVE-2026-21510), an authentication coercion flaw (CVE-2026-32202) remained. This gap between path resolution and trust verification left a zero-click credential theft vector via auto-parsed LNK files,” Akamai said in a Thursday report.

As Microsoft explains, remote attackers who successfully exploit the vulnerability in low-complexity attacks by sending “the victim a malicious file that the victim would have to execute,” could “view some sensitive information” on unpatched systems.

Microsoft flagged the CVE-2026-3220 flaw as exploited in attacks on Sunday after BleepingComputer reached out last week to ask why the advisory released during the April 2026 Patch Tuesday had an exploitability assessment of ‘Exploitation Detected’ while the vulnerability was flagged as not exploited.

A Microsoft spokesperson has yet to reply to a second email requesting more information about the CVE-2026-32202 attacks, including whether APT28 hackers also exploited this zero-click vulnerability.

Feds ordered to patch by May 12

On Tuesday, CISA added CVE-2026-32202 to its Known Exploited Vulnerabilities (KEV) Catalog, ordering Federal Civilian Executive Branch (FCEB) agencies to patch their Windows endpoints and servers within two weeks, by May 12, as mandated by Binding Operational Directive (BOD) 22-01.

“This type of vulnerability is a frequent attack vector for malicious cyber actors and poses significant risks to the federal enterprise,” the cybersecurity agency warned.

“Apply mitigations per vendor instructions, follow applicable BOD 22-01 guidance for cloud services, or discontinue use of the product if mitigations are unavailable.”

Although BOD 22-01 applies only to U.S. federal agencies, CISA has urged all security teams to prioritize deploying patches for CVE-2026-32202 and securing their organizations’ networks as soon as possible.

Threat actors are also actively exploiting three recently disclosed Windows security vulnerabilities (dubbed BlueHammer, RedSun, and UnDefend) in attacks aimed at gaining SYSTEM or elevated administrator privileges, with the latter two still awaiting patches.

AI chained four zero-days into one exploit that bypassed both renderer and OS sandboxes. A wave of new exploits is coming.

At the Autonomous Validation Summit (May 12 & 14), see how autonomous, context-rich validation finds what’s exploitable, proves controls hold, and closes the remediation loop.

Advertisement

Claim Your Spot

Food handling is an area of work that still relies heavily on humans. Fruit, vegetables, meat, and other foods need to be handled quickly but gently. It is also hard to automate because no two pieces of fruit, vegetables, or chicken nuggets look exactly the same.

Eka’s demos suggest that the company may be onto something big. I found myself mentally comparing their robots to GPT-1, OpenAI’s first large language model, developed four years before ChatGPT. GPT-1 was often incoherent but showed glimmers of general linguistic intelligence.

The robots I saw seem to have a similar kind of nascent physical intelligence. When I watched a video of one reaching for a set of keys in slow motion, I noticed it did something that seemed remarkably human: It touched the tips of its grippers to the table and slid them along the surface before making contact with the keys and securing them between its digits. Eka’s algorithms seem to know instinctively how to recover from a fumble. This kind of thing is difficult for other robots to learn, unless the humans training them deliberately make a wide range of mistakes.

Unlike with any other robot I can think of, it’s almost possible to imagine what the world is like for the robot. Its sensors seem to feel the weight of its arm, the inertia as it sweeps toward the keys and slows down. Once it has the keys in its grasp, it seems to sense the weight of them dangling from its claw.

I don’t know if Eka’s approach really is the route to a ChatGPT-like breakthrough in robotics. Some very smart experts believe that mixing human demonstration with simulation will yield better results than simulation alone. Maybe some combination of the two approaches will ultimately be necessary? But it does seem clear that robots will eventually need to have the kind of tactile, physical intelligence that Eka is working on if they are to obtain humanlike dexterity.

Agrawal tells me that the same general approach should work for finer manipulation. The fiddly dexterity required to build an iPhone, for instance, could be achieved by building different actuators and sensors and practicing the task in simulation.

After spending a few hours at Eka, I decide to stop by the restaurant downstairs. I watch from the counter as the staff prepare food and make coffee. A descendant of the machine upstairs may be able to do these things just as well, if not better. But given how much I enjoy chatting with the people who work there, I think I would pay extra to keep humans around. Unless, that is, my hands get automated away too.

Let us know what you think about this article. Submit a letter to the editor at [email protected].