gshc2020.com

An AI agent is a different kind of consumer than what most APIs were built for. Unlike a conventional application following a predefined workflow, an agent can decide at runtime which tools to use, sequence actions on its own and keep working toward a goal without a person checking every step.

So, I’ve started thinking about an API differently. It’s not just a gateway to data anymore; it’s an execution surface. Giving an agent access means giving that agent the ability to act on your behalf.

The Risk Teams Underestimate

Teams worry about an agent making an obvious bad call. What trips organizations up is more subtle: an agent making a call that’s completely valid from the system’s point of view and still produces an outcome nobody wanted.

Authentication, authorization and rate limits are still necessary, but not enough on their own anymore. An agent-ready platform also needs action-level permissions, current state, structured responses, protection against duplicate actions, complete audit trails and clear points where a human must approve something.

I see this in financial services, which is my field of experience and expertise. A “200 OK” response doesn’t tell the whole story. An event can execute successfully and still violate a restriction, duplicate an instruction already sent or process after a state change. For an agent, technical success and business correctness are two different things. Today’s APIs do not make that distinction explicitly.

The numbers back this up. A 2026 Salt Security survey of 327 security leaders found 48.9% of organizations lacked visibility into machine traffic and couldn’t monitor what agents were doing; 47% said API-security concerns had delayed a production release. Adoption is moving faster than governance, and I don’t think that gap closes on its own.

What I’ve Changed My Mind About

Previously, I would have told you good API design meant exposing functionality cleanly and letting developers build around it. I still believe that, but I now think it’s incomplete.

APIs should be designed backward from the business outcome, not outward from the underlying system function, communicating not just what action is available, but the current state, the conditions under which it’s permitted and how to recover if something goes wrong.

For a financial transaction, that might mean exposing transaction state explicitly, requiring an idempotency key so the same instruction can’t execute twice, and separating the permission to prepare an order from the permission to submit it. None of that is exotic, it’s the discipline that keeps a small mistake from becoming a real loss.

Documentation can’t be written for a human reader alone anymore, either. Schemas, permissions and what errors mean increasingly need to be structured and machine-readable, so an agent can interpret them correctly instead of guessing.

Where I Think The Industry Has It Wrong

I don’t think every API needs to be torn down and rebuilt for this. Agents don’t make REST (Representational State Transfer) obsolete, and adopting a new protocol doesn’t automatically make a platform agent ready. If an API is consistent, deterministic, well documented and returns structured data with meaningful errors, an agent should generally work with it as-is.

Where I get concerned is workflows that change financial or legal states, depend on unwritten business rules or produce ambiguous outcomes, places where a human has always filled the gaps with context and judgment. Handing an agent an endpoint doesn’t hand it that judgment.

What I’d Actually Prioritize

If I had a limited window to make a platform agent ready, I wouldn’t try to do everything at once. I’d inventory what agents could access, and classify those capabilities by consequence, read-only, reversible and actions carrying financial or regulatory weight, then pick one or two contained workflows where the value is obvious and the risk is manageable.

From there, I’d give every agent its own identity with narrowly scoped permissions, require human approval for anything high-impact and make sure the platform can confirm current state before an agent proceeds. I’d also invest in observability: audit logs, monitoring and a way to stop an agent immediately.

What I would not do is rebuild the entire API estate or launch a dozen use cases at once. That tends to create more risk than value.

The Cost Of Getting This Wrong

The consequences go beyond any single technical incident. If an API can’t clearly communicate permissions, the state and outcomes, clients build extra reconciliation work around it, which slows down and raises the cost of every integration. Some will absorb that cost; others will just pick a provider that gets them to production faster.

I’ve also seen companies invest heavily in agents and still need people reviewing routine actions manually, paying for the complexity of AI without getting the efficiency it was supposed to deliver. Trust ends up mattering more than uptime in deciding which platforms get automated against, and which get left out.

My takeaway is that agents don’t require abandoning solid API fundamentals. They require extending them. Platforms that treat identity, permissions and auditability as part of the transaction itself, not bolted on afterward, will be the ones agents, and the people relying on them, can trust to act.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?

In a study published in Nature Communications, researchers from the Institute of Industrial Science, The University of Tokyo, and collaborating institutions built optical structures inspired by the "Smith hat." This unusual shape is known for solving the so-called Einstein problem in mathematics. When the team illuminated the structures with laser light, they observed diffraction effects unlike those seen in conventional quasicrystals.

The Shape That Solved the Einstein Problem

The Einstein problem asks whether a single tile shape, known as a "monotile," can cover an entire surface without creating a repeating pattern.

Familiar tilings such as checkerboards and honeycombs repeat in a regular way. An aperiodic monotile, by contrast, can fill a surface without ever settling into a repeating arrangement.

In 2023, researchers discovered the first such monotile, the Smith hat. The finding attracted widespread attention because it provided a long-sought solution to the mathematical problem.

"What is especially fascinating about the hat tile is that, although the resulting pattern appears irregular at first glance, it is actually constructed from the honeycomb lattice," says lead author Yuto Moritake. "We wanted to see whether this unique shape could also produce any unexpected physical phenomena."

Turning a Mathematical Pattern Into an Optical Structure

To test that possibility, the researchers created nanoscale versions of the pattern on silicon nitride films using electron beam lithography.

When laser light was directed at the structures, the resulting diffraction patterns formed distinctive pinwheel-like shapes. These patterns directly revealed the chiral character of the aperiodic structure.

Chirality refers to a form of handedness in which a structure and its mirror image cannot be perfectly matched. In this case, the unusual arrangement of the monotile pattern caused the light itself to display a chiral response.

"We found that the diffraction patterns themselves become chiral because the structure lacks mirror symmetry," explains senior author Masaya Notomi. "This kind of optical response is fundamentally different from that observed in conventional quasicrystalline materials."

Light Responds to Direction and Polarization

The researchers also found that the diffraction pattern changed depending on both the direction and polarization of the incoming light.

When the physical structures were mirrored, their optical behavior reversed as well. This showed that the response of the light was directly tied to the symmetry of the underlying pattern, revealing a new type of symmetry-controlled optical behavior.

"These results open a new direction of research on the fusion of quasiperiodic order and chirality," remarks Moritake. "Monotile patterns provide a platform for exploring optical phenomena that emerge from the interplay of symmetry, chirality, and aperiodicity."

From Abstract Mathematics to New Optical Physics

The researchers say structures inspired by monotiles could eventually contribute to technologies designed to manipulate light, control polarization, and support advanced optical devices.

More broadly, the findings show how a discovery that began as an abstract mathematical puzzle can lead to unexpected physical effects. The Smith hat was first celebrated for solving a question about how shapes can cover a surface. Now, its unusual geometry may also help researchers uncover new ways to control and study light. %!s()

Plus, the festival said that it would wrap up its multi-year program “Dead Angle” with a focus on the theme Common Ground, following Borders in 2024 and Institutions in 2025.

IDFA said that its 2026 edition will also feature Huezo’s personal top 10 film selection and a talk featuring the filmmaker. Born in El Salvador and raised in Mexico, she “has developed a distinctive body of work that moves between documentary and fiction, exploring how the violent consequences of organized crime and political turmoil shape the intimate lives of local families and communities,” fest organizers said. “Her cinema is rooted in close observation and shaped by a highly attentive visual and sonic language – allowing memories and experience to surface through atmosphere and deeply personal stories.”

Huezo first gained international recognition with her feature-length debut The Tiniest Place (2011), a portrait of a remote Salvadoran village whose inhabitants carry the memory of a massacre during the country’s civil war. Her fiction debut, Prayers for the Stolen (2021), is a coming-of-age story about girls growing up in the mountains of Mexico under the threat of human trafficking by local drug cartels.

The spotlight on Greaves (1926–2014) is designed to bring “a rediscovered voice in cinema back into view,” IDFA said. Across a formally inventive filmography shaped by sharp observation of social and political life, Greaves explored Black history and culture.”

The spotlight program includes Once Upon a Time in Harlem (2026), capturing a 1972 reunion of key Harlem Renaissance figures in Duke Ellington’s apartment, completed posthumously by Greaves’ son and granddaughter David and Liani Greaves. Also featured are Symbiopsychotaxiplasm: Take One (1971), which turns a fiction shoot in Central Park into a self-reflexive study of a film crew in revolt, and a director’s cut of The Fight.

About the Dead Angle: Common Ground focus, meanwhile, will be “exploring the many ways people come together – from extraordinary acts of solidarity and collective resistance to the everyday life of communities and quintessential function of public space,” IDFA highlighted. Among confirmed titles exploring solidarity in times of crisis are Everybody to Kenmure Street (2026) by Felipe Bustos Sierra, Shinsuke Ogawa’s The Battle Front for the Liberation of Japan – Summer in Sanrizuka (1968) and Barbara Kopple’s Oscar-winning American Dream (1990).

Finally, IDFA DocLab returns this year with the exhibition of what the fest described as “a diverse lineup of interdisciplinary immersive art and anniversary highlights to be revealed in the coming weeks.” %!s()

Researchers found that the energy associated with these winds is about 100 times greater than earlier estimates, revealing that black holes can affect vast regions of space on a truly enormous scale.

"Black holes are largely known for sucking matter in, but they also eject gas in the form of powerful winds," says Satoshi Yamada, Assistant Professor at Tohoku University's Frontier Institute for Interdisciplinary Sciences (FRIS). "These winds were thought to be contained within the galaxy, but our study revealed that the force is immensely more powerful than previously understood."

XRISM Takes a Closer Look at a Distant Quasar

Yamada and his colleagues, including researchers from Kanazawa University and Tokyo Metropolitan University, among others, used the X ray astronomy satellite XRISM to study the quasar H1821+643. The object lies in the constellation Draco, about 3.4 billion light-years from Earth.

The rapidly growing black hole at the center of H1821+643 sits within a galaxy cluster. There, it stirs the surrounding hot gas, which gives off X-rays. To measure how that gas was moving, the researchers analyzed emission lines produced by iron ions.

Black Hole Turbulence Reaches Far Beyond the Galaxy

XRISM's high-precision measurements showed that the extremely hot gas around the black hole is far from motionless. Instead, turbulence causes it to spread violently across a broad region.

The team also found that this gas flow extends beyond the black hole's host galaxy, reaching distances of about 300,000 light-years. Even more striking, the energy contained in the turbulence was approximately 100 times greater than previous estimates.

That amount of energy is comparable to the output of several billion supernova explosions, the powerful blasts produced when certain stars reach the ends of their lives.

"For the first time, we have shown that black holes influence the broader cosmic environment through a shock wave of astonishing power," adds Yamada. "Black holes are key drivers of gas flows and motion in space, transporting vast amounts of energy to different regions of the cosmos."

A Bigger Role for Black Holes in the Cosmos

The findings suggest that supermassive black holes do more than consume nearby matter. They can also send energy and gas far into surrounding space, potentially helping shape the environments around galaxies and galaxy clusters.

Details of the research were published in the journal Nature Astronomy.

Future observations are expected to provide a clearer picture of how black holes affect their surroundings and how matter and elements move through different regions of the cosmos. %!s()

Still, people rarely want to take the same class twice, hence why padding-compacted eyebrows have been raised by Alyx again being a tentpole of the new Steam Frame headset – this time, to showcase the game running in PC-less, standalone VR. Why not make something new, even a smaller tech demo like ADJ or the Index’s own The Lab collection of Portal-themed minigames?

According to Valve staff, several of whom I spoke to prior to the Frame’s Monday night launch, there were several reasons why a return to Alyx was considered more valuable – even though a brand new VR game was, at least briefly, on the table.

"We had talked about things like that," said SteamOS developer Pierre-Loup Griffais. "We decided to focus on just making sure the Steam catalogue as a whole would be a good experience, and trying to put most of our efforts into that compatibility. Because there's a tonne of existing experiences I think that can intro you to whatever side of the devices you're more interested in, right?"

Griffais refers to a key Steam Frame party trick: the ability to run both VR and flat games, from an existing Steam library, in the manner of your choosing, be it over conventional PC VR wireless streaming or standalone via the headset’s own internal hardware. "I mean, I can't speak for everyone here, but if I'm showing the headset to someone, something like SuperHot VR or Beat Saber or one of those OG titles is gonna get on there," he contined. "Because they're still so good, right? And I think there's a tonne of that catalogue that can actually serve as a good launch experience for most people because they maybe haven't played them yet."

Robin Walker, a Valve programmer and co-creator of Team Fortress 2, offers a different view: that the process of re-developing Alyx to both run and perform well on the Frame’s dinky processing parts, with all the software rejiggery that entailed, provided Valve with invaluable knowledge on how to get Frame games working better in general.

"When you're building a new piece of hardware and there's a whole bunch of new software in that software stack, and there's drivers and all that sort of stuff, having a clear target is incredibly useful for you to make the tradeoffs," Walker explained. "To be able to test, to experiment, try different approaches, and Alyx was a far, far more useful target for us there than The Lab would have been, just because of the goals of the product."

"It was so worth it," added engineer Autumn Aston, "to just do that work anyway because we found so much stuff. So many, y’know, bugs, performance things in the driver. It was like a great way to get everything like up to scratch for the whole [Frame] project."

"All the fixes we got in for Alyx, the vast majority of them are applicable to basically anything that's running on the Frame," said Walker. Valve have since been sharing this knowledge with other VR game developers as part of Steam Frame Verified programme: a system, practically identical to previous schemes for the Steam Deck and Steam Machine, where developers can get their games rated for compatibility with the Frame’s standalone mode, and receive help from Valve directly on getting their games more playable on the headset.

Ironically, there was a time – likely the same as when Valve were considering a new VR game – when Alyx’s own developers didn’t think it possible that such a game could run adequately on a Snapdragon smartphone CPU. It was, so says Walker, the Frame’s own developments in performance-saving tracking tech – and a healthy supply of RAM - that made standalone Alyx a possibility.

"I think at the outset, as a team, when we thought about... it's maybe going back a year and a half, two years ago, when we started thinking about Alyx running natively on the Steam Frame. It just seemed faintly ludicrous to us," he recalled. "Alyx was not a product that was designed with really any compromises around performance or anything.

"But then, about a year and a half ago, Autumn started doing prototyping. And the specific things that gave us hope there, I guess, or the things that made us think, 'Hey, maybe there's something interesting here' was the foveated rendering and the eye tracking, because that seemed like potentially a silver bullet that could allow us to run this thing natively on the headset. And then the other bit also was the 16GB of RAM, which meant we felt like there was a chance there. If that had been 12GB or 8GB, there was not even really. It was just it would be much less interesting to us."

Also ironic: that memory, which is partly why the £889-plus Frame costs as much as it does, was allegedly also instrumental in making one of its best available games run using one of its most unique features. Would I rather trade all of that an new Valve game? Probably yes, because Valve games are good. But if that’s not happening, I suppose at least Frame owners will get better standalone games out of it. %!s()