Connect with us

Tech

How AI changes the math for startups, according to a Microsoft VP

For 24 years, Microsoft’s Amanda Silver has been working to help developers — and in the last few years, that’s meant building tools for AI. After a long stretch on GitHub Copilot, Silver is now a corporate vice president at Microsoft’s CoreAI division, where she works on tools for deploying apps and agentic systems within enterprises.

Her work is focused on the Foundry system inside Azure, which is designed as a unified AI portal for enterprises, giving her a close view of how companies are actually using these systems and where deployments end up falling short.

I spoke with Silver about the current capabilities of enterprise agents, and why she believes this is the biggest opportunity for startups since the public cloud.

This interview was edited for length and clarity.

So, your work focuses on Microsoft products for outside developers — often startups that aren’t otherwise focused on AI. How do you see AI impacting those companies?

I see this as being a watershed moment for startups as profound as the move to the public cloud. If you think about it, the cloud had a huge impact for startups because it meant that they no longer needed to have the real estate space to host their racks, and they didn’t need to spend as much money on the capital infusion of getting the hardware to be hosted in their labs and things like that. Everything became cheaper. Now agentic AI is going to kind of continue to reduce the overall cost of software operations again, because many of the jobs involved in standing up a new venture — whether it’s support people, legal investigations — a lot of it can be done faster and cheaper with AI agents. I think that’s going to lead to more ventures and more startups launching. And then we’re going to see higher-valuation startups with fewer people at the helm. And I think that that’s an exciting world. 

What does that look like in practice?

Techcrunch event

Boston, MA
|
June 23, 2026

We are certainly seeing multistep agents becoming very broadly used across all different kinds of coding tasks, right? Just as an example, one thing developers have to do to maintain a codebase is stay current with the latest versions of the libraries that it has a dependency on. You might have a dependency on an older version of the dot-net runtime or the Java SDK. And we can have these agentic systems reason over your entire codebase and bring it up to date much more easily, with maybe a 70% or 80% reduction of the time it takes. And it really has to be a deployed multistep agent to do that.

Live-site operations is another one — if you think of maintaining a website or a service and something goes wrong, there’s a thud in the night, and somebody has to be on call to get woken up to go respond to the incident. We still do have people on call 24/7, just in case the service goes down. But it used to be a really loathed job because you’d get woken up fairly often for these minor incidents. And we’ve now built a genetic system to successfully diagnose and in many cases fully mitigate issues that come up in these live site operations so that humans don’t have to be woken up in the middle of the night and groggily go to their terminals and try to diagnose what’s going on. And that also helps us dramatically reduce the average time it takes for an incident to be resolved.

One of the other puzzles of this present moment is that agentic deployments haven’t happened quite as fast as we expected even six months ago. I’m curious why you think that is.

If you think about the people who are building agents, what is preventing them from being successful, in many cases, it comes down to not really knowing what the purpose of the agent should be. There’s a culture change that has to happen in how people build these systems. What is the business use case that they are trying to solve for? What are they trying to achieve? You need to be very clear-eyed about what the definition of success is for this agent. And you need to think, what is the data that I’m giving to the agent so that it can reason over how to go accomplish this particular task?

We see those things as the bigger stumbling blocks, more than the general uncertainty of letting agents get deployed. Anybody who goes and looks at these systems sees the return on investment.

You mention the general uncertainty, which I think feels like a big blocker from the outside. Why do you see it as less of a problem in practice?

First of all, I think that it’s going to be very common that agentic systems have human-in-the-loop scenarios. Think about something like a package return. It used to be that you would have a workflow for the return processing that was 90% automated and 10% human intervention, where somebody would have to go look at the package and have to make a judgment call as to how damaged the package was before they would decide to accept the return. 

That’s a perfect example where actually now the computer vision models are getting so good that in many cases, we don’t need to have as much human oversight over inspecting the package and making that determination. There will still be some cases that are borderline, where maybe the computer vision is not yet good enough to make a call, and maybe there’s an escalation. It’s kind of like, how often do you need to call in the manager? 

There are some things that will always need some kind of human oversight, because they’re such critical operations. Think about incurring a contractual legal obligation, or deploying code into a production codebase that could potentially affect the reliability of your systems. But even then, there’s the question of how far we could get in automating the rest of the process.

source

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Tech

Amazon Music partners with Bandsintown for concert listings

Amazon Music said today that it is partnering with live event listing platform Bandsintown to bring concert listings to its music streaming service.

Through a Bandsintown for Artists integration, users will be able to find live shows on the artists’ profile pages on Amazon Music. They can also click on the “buy ticket” buttons to purchase tickets on Bandsintown. (Artists will need to link their Amazon Music profile for the first time on Bandsintown for event sync to begin.)

What’s more, Amazon Music will automatically list events from venues, festivals, and promoters using Bandsintown Pro on its platform.

“Live music is one of the most powerful ways fans connect with the artists they love—something we’ve seen firsthand through the livestream performances we’ve brought to fans worldwide,” said Karolina Joynathsing, Director of Business Development, Amazon Music, in a statement. “That’s why we’re thrilled to team up with Bandsintown, so fans can now discover upcoming shows from the artists they love on Amazon Music, alongside streaming their music, catching exclusive livestreams, shopping merch, and more, all in one place,” she said.

The feature is rolling out now and will be fully available on Amazon Music this spring on both iOS and Android apps, the company notes.

Bandsintown said that it has more than 700,000 artists on the platform, along with over 65,000 venues and festivals using its Bandsintown Pro marketing and promotion product. It added that more than 100 million users are registered on the platform, but didn’t specify the number of active users.

Previously, Bandsintown had partnered with Amazon Music in 2023 to let artists sell merchandise on the music streaming platform. Amazon Music is relatively late to the concert discovery. Spotify has been working on it for years and Apple Music teamed up with Ticketmaster and Bandsintown to power its live events feature in March. Last year, Soundcloud also worked with Ticketmaster to let artists list their live events.

Techcrunch event

San Francisco, CA
|
October 13-15, 2026

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

source

Continue Reading

Tech

AI is spitting out more potential drugs than ever. This start-up wants to figure out which ones matter.

AI’s biggest impact in science is Google DeepMind’s use of a deep learning model to predict the complex structures of proteins — the molecules that drive virtually every process in living cells.

But as AI models continue to spit out more candidates for potential treatments, there’s an emerging bottleneck: actually characterizing all those candidates in practice, for testing and mass production.

That’s the goal of 10x Science, a startup founded in December 2025 that announced a $4.8 million seed round today, led by Initialized Capital and with backing from Y Combinator, Civilization Ventures, and Founder Factor. Its three founders are David Roberts and Andrew Reiter, experienced biochemists, and Vishnu Tejas, a serial founder with expertise in computer science and AI models.

“When biopharma tries to create a drug candidate, they have all of these really nice prediction tools,” Roberts told TechCrunch. “You can add as many candidates as you want to the top of the funnel, but they all have to pass through this characterization process. Everything needs to be measured.”

Understanding the structure of proteins is key for researchers developing biologic drugs, which are produced in living cells and use sophisticated design to specifically target diseases and conditions. For example, they can be designed to target specific cells, like Keytruda, a popular drug sold by Merck that helps the immune system identify and attack cancers.

10x’s three founders worked together in the Stanford lab of Nobel laureate Dr. Carolyn Bertozzi, where they studied the interactions between cancer cells and the immune system, and were frustrated by their inability to understand precisely what was happening on a molecular level.

The most accurate way to assess molecules is through a complex technique called mass spectrometry, a way of determining their atomic structure by measuring them in an electric field. The relatively new technique generates complex data that requires significant expertise to interpret, and analyzing it takes up a lot of time.

Techcrunch event

San Francisco, CA
|
October 13-15, 2026

10x’s platform combines deterministic algorithms rooted in chemistry and biology with AI agents that can interpret that data. The team had to do significant work to train the models on spectrometry data and make its analyses traceable, a key requirement for a tool that will be used to help companies achieve regulatory compliance.

Matthew Crawford is a scientist at Rilas Technologies, a firm that runs chemical analyses for other companies — saving clients like biotech startups from having to invest several million dollars in their own spectrometry equipment and the experts to operate it. Crawford has been using the 10x Science platform for several weeks and says it is speeding up his work.

Crawford said the model surprised him with its ability to explain its conclusions, find the right data for analyses on its own, and adapt to evaluating different kinds of molecules. While some AI tools he has experimented with in the past over-promised or suffered accuracy issues, he says this one makes reasonable assumptions, something he attributes to the deep domain expertise of its creators.

“I ran a particular protein through it, and it just kind of figured out, from what I named the file, what the protein probably was,” Crawford said. “It then searched databases online for the sequence for that protein, so I didn’t have to program in the sequence.”

10x executives say they’re also working with multiple major pharmaceutical companies, as well as academic researchers. The plan is to use this seed funding to hire more engineers and continue to refine the model and offer it to new customers. If they are able to gain traction characterizing proteins, Roberts hopes the company will expand to offer a new kind of understanding of biology, combining protein structure with other data about cells.

“The deeper thing behind what we’re building is actually a new way to define molecular intelligence,” Roberts said.

For its investors, 10x offers a useful way into the biotech space that isn’t dependent on a specific drug succeeding and winning regulatory approval. If the company works out the way its founders hope, it will become an important tool for drug development, whether or not the eventual products succeed in the marketplace.

“This is a SaaS platform that pharma has to pay for, every single month, to go through all of these potential candidates,” Zoe Perret, a partner at Initialized, said. She’s counting on the deep experience of the founders to protect the company from competitors; there simply aren’t that many people who understand these methods and the data they produce.

What the platform could do, Crawford says, is help unlock the techniques for researchers who could benefit from these methods but lack the time or resources to deploy them.

“Groups here are trying to make a new drug,” he told TechCrunch. “They just want to get a quick, simple answer out of mass spec, and then it opens up a whole can of worms. This software is going to help keep that can of worms closed and just get them the answer they actually need to then do the next thing in their research.”

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

source

Continue Reading

Tech

Snap Map’s new ‘Place Loyalty’ badges will show the spots you visit most often

Snapchat is rolling out a new “Place Loyalty” feature that will show users when they’re among the most frequent visitors to a place on Snap Map over the past year, the company announced on Wednesday.

If you’re in the top 25% of users who visit a certain place, you’ll see your ranking on the Snap Map and can choose to share it with others. You’ll receive a gold badge if you’re among the top 1% of visitors, silver if you’re top 10% and bronze if you’re top 25%. For brands and chains, Snapchat aggregates visits across all locations.

Of course, you will only see these badges if you choose to share your location with Snapchat. Additionally, Place Loyalty rankings are only visible to you, Snapchat says.

The new feature is one more way for users to engage with the app and encourage them to share their loyalty badges on other social platforms or with friends in their social circle.

Image Credits:Snapchat /

Launched in 2017, Snap Map was originally seen as a way for users to see their friends’ locations and browse public snaps from around the world. Over time, the feature has evolved and now offers ways for users to discover local hotspots and find things to do. With this new feature, Snapchat is giving users another way to interact with the map.

Snapchat announced last year that Snap Map has surpassed 400 million monthly active users.

Snap Map originally gave the social network a competitive edge over rivals like TikTok and Instagram, but the latter took on Snapchat with its own “Instagram Map” last year. Instagram hasn’t shared any stats regarding how many people use the feature, so it’s unknown how exactly it stacks up against Snap Map.

Techcrunch event

San Francisco, CA
|
October 13-15, 2026

Given that Snap Map is one of Snapchat’s core products, the social media giant often updates it with new features to attract and retain users. In 2025, the company launched a “Footsteps” feature that lets you see how much of the world you’ve explored and track where you’ve traveled.

Snapchat also rolled out “Promoted Places” on Snap Map last year as a way for users to discover new places. For brands, Promoted Places offers the ability to advertise and showcase all of their locations on Snap Map.

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

source

Continue Reading