In 2012, T-Mobile was losing hundreds of thousands of customers a quarter. AT&T’s attempted acquisition of the Bellevue, Wash.-based wireless company had just collapsed. T-Mobile was the fourth-place carrier in a four-carrier market, with a network that was a punchline and no obvious plan for fixing any of it.

Mike Katz was one of the leaders responsible for figuring it out.

The result of their work was the “Un-carrier,” a series of moves starting in 2013 that upended one industry tradition after another, launched under CEO John Legere and continued by his successor, Mike Sievert.

T-Mobile today is worth nearly $195 billion and claims the country’s best network. The company is now led by Srini Gopalan, who became CEO in November 2025.

Katz, whose wireless career began selling VoiceStream phones at a Circuit City in Fort Collins, Colo., is leaving after 28 years. The company announced July 7 that Katz, then chief business and product officer, would pursue “new professional interests,” staying on as an advisor through December.

During his tenure at T-Mobile, Katz worked in sales and corporate strategy, ran the company’s prepaid business, led consumer marketing through the first Un-carrier moves, built the company’s business division, and served as chief marketing officer, before becoming chief business and product officer in December 2025.

In an interview with GeekWire, Katz reflected on his experience inside one of the most improbable and irreverent comebacks in American business.

Continue reading for the lessons we took from his story.

Use speed to your advantage against bigger competitors.

When the AT&T deal died, Katz was put on a small team assigned to assess the company’s position and come up with a plan. Their advantage, he said, was that T-Mobile was too small to compete on everyone else’s terms.

“We were so subscale relative to AT&T and Verizon at the start, and what we developed because of being subscale was agility,” Katz said. “We would do things, and AT&T and Verizon would take forever to respond, and by the time they responded, all their customers were gone.”

It started with Un-carrier 1.0. Under the standard model then, a customer could get a phone for well below cost if they signed up for a two-year plan. But the carrier recovered it through the monthly service rate, which never came down once the phone was paid off. Customers paid a penalty for leaving early.

T-Mobile separated the device from the service plan in March 2013, pricing service on its own and putting the phone on monthly installments that ended when the device was paid for.

Verizon didn’t stop signing new customers to two-year contracts until August 2015. AT&T held out until January 2016.

Getting copied was the point. Katz said one measure of success for every Un-carrier move was whether AT&T and Verizon would eventually imitate it, because by the time they did, T-Mobile had already gotten all the credit and attention.

Legere reveled in calling AT&T and Verizon “Dumb and Dumber” (it was never clear which was which in his eyes). He crashed AT&T’s party at CES in 2014 and was escorted out by security, generating multiple rounds of positive publicity for T-Mobile. He asked the public to vote on how to taunt his rivals next, with skywriting over their headquarters among the options.

They were clearly having fun. Asked whether T-Mobile ever worried it had gone too far in baiting its rivals, Katz said, “Not really.” The mission then, he said, was to change wireless for everyone and not just for T-Mobile’s own customers.

The mockery extended to the events themselves. Katz recalled preparing for an early Un-carrier launch in New York when Legere, around 10 p.m. the night before, decided he wanted to open by parodying AT&T’s “It’s Not Complicated” campaign, in which a man interviewed small children at a tiny table. That meant finding dolls. Katz said the team spent the night working out how to get into the American Girl store in Manhattan.

“It was wild,” Katz said. “There were so many things in those early events.”

The next morning the lights came up on Legere at a small table, talking to the dolls.

“Maybe we were the only ones that got it,” Katz said. “But it was funny to us.”

Look for the unexpected and find out why it’s happening.

Back in 2010, then-CEO Robert Dotson put Katz in charge of the company’s prepaid business. It was not a marquee assignment.

Prepaid customers of that era “didn’t really pick prepaid. Prepaid picked them,” Katz said. “They had no other choices. They had tough credit, or they socioeconomically were in a really tough place.”

T-Mobile was also losing ground. A price war broke out in January 2009, when Sprint’s Boost Mobile introduced a $50 unlimited plan and MetroPCS answered at $40. The recession was pushing customers toward prepaid, and the flat-rate carriers were taking market share.

Katz ran prepaid as a business of its own, operations and marketing together, and looked for a way to compete without matching the price. Those cheaper plans limited customers to 3G, so T-Mobile charged a bit more, gave them 4G, and called it Monthly 4G. (A deliberately clear and functional name, as he pointed out.)

It worked. Through 2010, 2011 and 2012, while the postpaid business was shedding customers, prepaid was the only part of T-Mobile that was growing.

The surprise: a lot of the new growth was coming from previously postpaid customers, people who had other options and were choosing prepaid anyway. Katz’s team asked them why.

“There’s no contract. I sign up for a $50 plan, and it actually costs $50 a month. It’s just simple and predictable,” Katz said, quoting customers. It might seem obvious now, he said, “but at the time it’s like, wow, that’s a pretty interesting insight.”

Expand what’s working to other parts of the company.

They found inspiration in that when they turned to what T-Mobile should do next. The idea was to give postpaid customers the same simplicity that was drawing them to prepaid.

They called it pain-free wireless. They mapped out six moves in advance, including the end of upgrade restrictions and international roaming fees. These were the origins of the Un-carrier.

They pitched it to Legere, and to Sievert, the chief marketing officer at the time. Both signed on. The launch was originally set for the end of 2013. Legere moved it to the beginning of the year.

“There is something about desperation that really helps create crisp decision-making,” Katz said. “You can keep trying to make a gameplan that’s been failing for years work, or take some swings and take on some risk. … Because what’s the downside? It can’t get much worse than it was.”

Make sure people know the ad was yours.

Storytelling is the thing companies and brands get wrong most often, Katz said, and he called it the biggest thing he learned from Sievert, whom he worked with for more than a decade. It’s especially important in the wireless industry, where customers are effectively buying a promise.

“At the end of the day, we sell invisible air. You don’t really see the product we sell,” Katz said. “There’s phones, but we don’t make the phones. Apple makes the phones. There’s towers that you can see, but the tower companies, those are their towers. We sell the invisible air in between.”

Katz divides the profession into two camps: the “award show CMOs,” motivated by collecting trophies at Cannes, and the ones who measure themselves by business impact. He puts himself in the second group.

One example: T-Mobile’s February 2025 Super Bowl spot, which launched the public beta of its Starlink partnership, offering satellite-to-cell texting in the parts of the country terrestrial networks don’t reach.

It was tempting to fill it with celebrities and gimmicks, he said. Instead the spot was built around a voiceover and a direct call to action: try it out for free during the beta, open to AT&T and Verizon customers too.

“If we wanted to build a Super Bowl commercial to win the best Super Bowl spot in Ad Age, we would have done something very different,” Katz said. “We would have done the celebrity-palooza.”

“Within 30 minutes of that Super Bowl spot running, we achieved all the goals that we had for the rest of the month in signups,” Katz said.

He contrasted that with Verizon’s 2024 Super Bowl ad starring Beyonce, which he said people recalled without necessarily remembering the advertiser.

Bad results don’t mean bad people.

One of the lessons Katz learned from Legere, in addition to taking risks and failing fast, came from how the incoming T-Mobile CEO treated the people who were already there when he arrived. Legere took over a company that was losing badly and kept the team in place.

“It would have been really easy for him to come in and assume, hey, company’s not doing well, we must have a bunch of bad people here, and change everybody out, and he didn’t do that,” Katz said. “Don’t assume simply because the results of the company are bad that the people are bad.”

Katz said that was in his head during the Sprint merger, where the same thing applied: a business in rough shape, staffed by people who weren’t the reason. T-Mobile’s leadership today is a mix of the two companies.

Perception can sometimes lag reality.

Katz said T-Mobile passed Verizon on network quality years before customers believed it. Verizon “had so much built up brand equity around being the best network,” he said, that four or five years later, people still named Verizon when asked who had the best network.

T-Mobile declared victory at an event in Bellevue in June 2025, citing an Ookla study built on half a billion crowdsourced data points. Verizon disputed the methodology, saying crowdsourcing can’t control for variables, and pointed instead to RootMetrics drive tests that it said still showed Verizon the most reliable.

T-Mobile then began running ads with Billy Bob Thornton declaring it the best network in the country. Earlier claims had carried qualifiers, like fastest 5G or most available 5G. This one didn’t. Katz said the gap in how customers rate the two networks is now down to a couple of points.

Staying on top is harder than the turnaround.

For all the attention the comeback gets, Katz said the more remarkable part has been what came after it.

“If you look at companies that have done big turnarounds, you don’t really see a lot that have had the kind of sustained success and continued growth for a decade and a half like T-Mobile has,” he said.

He gives much of the credit to the network. The Sprint merger gave T-Mobile a spectrum position it didn’t have before, and the company bet that 5G would play out in mid-band spectrum, which was seen as risky at the time. Protecting that advantage through later spectrum deals was one of the things he worked on.

Building the network was not.

“There’s a couple jobs that they never let me do, probably for very good reason,” Katz joked. “You actually need to know what you’re doing.”

A mature company also needs different things from its leadership than a struggling one does. Gopalan spent years on T-Mobile’s board before joining as COO and then taking over as CEO. Katz describes him as an operator who brings discipline to the details of a business that by now has years of success behind it.

Asked for his biggest mistake, Katz said he moved too slowly at times.

“There are big decisions, risky decisions, that I could have made a lot faster,” he said. “I wish I’d pushed harder to make some of those decisions.” Some of what became the early Un-carrier moves were ideas that had been floated years earlier, in corporate strategy.

Katz says he still believes the company’s best days are ahead.

As for what’s next for him? Stay tuned, he says.

OpenAI’s new Astra model will use “recurrent depth,” a technique that allows the model to operate outside of the sequential thinking that characterizes most reasoning models.

The once-dominant mapping app has surged to the top of Apple’s U.S. App Store after refusing to adopt Trump’s “Lake America” name, drawing more than half of its 2026 U.S. downloads in just six days.

OpenAI CEO Sam Altman said AI adoption is "non-negotiable" for countries, comparing rejecting it to refusing electricity a century ago and predicting it will unleash an unprecedented boom in entrepreneurship. "The economic growth and benefit to people that can come from this, the value to a country, is too high to ignore," he said. His comments were made during a fireside chat with U.S. Commerce Secretary Howard Lutnick at the G20 Innovation Ministerial in Chapel Hill, North Carolina. CNBC reports: Altman thinks the public won't be talking as much about AI a decade from now -- it'll be expected everywhere. "A kid growing up today will never be smarter than AI, but he or she will also never have understood a world where every product and service that they interact with is not really smart and really capable and really helpful," he said. Altman referenced the adoption of electricity multiple times in his comments and drew a line between that and AI. "I think it would be approximately as bad of an idea to say we're not going to have AI in our country as it was to say we're not going to have electricity in our country, you know, back 100 plus years ago," he said. Altman called cybersecurity one of the biggest challenges to navigate in the age of AI and one that leaders shouldn't sidestep. "I think some things are going to go very wrong with cybersecurity unless people act quite urgently," he said. Altman said that falling short in cyber defense and other areas "could set this technology back a great deal." You can watch a recording of the chat on YouTube.

Read more of this story at Slashdot.

It might be overwhelming to see all of the new vocabulary popping up in software development these days thanks to AI tools introducing them… all the time.

Some of this new vocab describes useful patterns that people are newly pursuing, others are just fancy names on top of things that already exist, and some are still actively being defined as we speak.

In our latest episode of the GitHub Podcast, Marlene Mhangami, GPS, and I talked through some of the AI terms developers are learning right now: loop engineering, Ralph loops, squads, harness engineering, hill climbing, forward deployed engineers, closed models, open weights, and open source models.

If you’re a reader instead of a listener, here’s a guide to what those terms mean, why they matter, and how to think about them.

Listen to the full episode below! 👇

Loop engineering: Moving beyond one-shot prompts

Loop engineering is the practice of designing repeatable systems around agents, instead of manually prompting them for one task at a time.

A simple example: instead of asking an agent every morning to review new issues, summarize them, and propose fixes, you create a loop that runs on a schedule. That loop might fetch issues, pass them to an agent, validate the output, and escalate anything that gets stuck. It’s a glorified AI-native cron job.

Ralph loops: The brute-force cousin of loop engineering

A Ralph loop is one implementation of this “loop” concept: you give an agent a detailed task, often from a product requirements document or spec, and have it keep working until the job is done.

That can be useful, especially for breaking down large tasks into repeated plan-act-check cycles. But, on the other hand, it can also be expensive and inefficient because every iteration uses more tokens, more context, and more compute.

Loop engineering aims to make this pattern more structured, so you’re not caught asking an agent to “try again” all the time. A well-designed loop adds primitives like skills, observability, validation, routing, and checkpoints.

Squads, fleets, and multi-agent workflows

If loops define a workflow, “squads” and “fleets” describe how multiple agents can participate in that workflow.

A squad is a group of agents with different roles. They often reflect a real-world team. One agent might plan, another agent might vet that plan, another agent might implement it, another might test it, and another might review it.

A fleet refers to parallel agents working on tasks at the same time. You can have a squad working in a fleet in parallel, or in a sequence.

Operating this way lets different agents handle different parts of a process, and you can fine-tune and specialize each one with specific skills to be more efficient.

The core idea is parallelization and specialization. Instead of one agent trying to do everything, different agents can handle different parts of a development process.

Harnesses: The system around the model

Outside of what a model generates, a harness is everything surrounding it that makes it useful in your workflows.

That could be the tools, permissions, memory, context, orchestration (and so on) that guides how the model behaves. If it helps you remember: harnesses are aptly named after the harnesses for horses. Horses are like models that can run wild, and a harness helps direct the horse’s weight safely as it completes tasks. Get it?

Anyway, a good example of a software harness is GitHub Copilot. It connects models to codebases, editors, pull requests, terminals, and so on.

When you hear the term “harness engineering” tossed around, that’s the work of designing and improving that system that surrounds the models.

Hill climbing: Improving agents with feedback

The term “hill climbing” is used to describe the process of improving agents and harnesses over time.

That could mean, for example, using evals to measure whether an agent is producing the right kind of output (and then adjusting the harnesses until the results improve).

Or, another example, if your agent is supposed to review pull requests, hill climbing might be checking if it indeed finds meaningful bugs and produces useful recommendations, and adjusting tooling to improve that.

Forward deployed engineer: A familiar role with an AI focus

A forward-deployed engineer job has already existed, but AI branding makes it sound edgy and new. Now, it’s a customer-facing software engineer, or sales engineer, or solutions engineer, often with an AI focus.

If you haven’t seen those job titles before, this person generally works closely with customers to implement or adapt technical solutions into their environments. With the AI focus, that means helping teams integrate AI tools, workflows, agents, etc. into their existing systems.

Closed models, open weights, and open source models

Not all models are shared in the same way.

Closed models are accessed through an API or hosted product. Developers can use the model, but they don’t get access to the underlying weights, training data, or training process. The big, famous frontier models you hear about are often all closed models.

Open weight models make the model weights (which are like dials that decide how important certain inputs are) available. Developers can download and run these models, often locally or in their own infrastructure. But, to be clear, the dataset and training method may not be fully available.

Open source models go a step further, in that the model, code, data, and training process are all available for inspection, reuse, and modification.

The more open the model, the more you can run, customize, audit, and trust it.

The terms are ever-evolving

This is just a sampler of some of the terms we’re hearing a lot today. Some will stick around, and others will fade into our memories, and others will be replaced by better language as the industry matures.

Don’t worry about falling behind on buzzwords. They’re just words, and more important are the practices under them! Ask yourself if workflows can repeat reliably, how you validate tasks, how humans should (or shouldn’t) interfere, how much you can rely on a model, and how you can improve that your system. It’s a new era of engineering, and best practices still matter!

The post Decoding the new AI lingo: Loops, harnesses, squads, hill climbing… oh my! appeared first on The GitHub Blog.

Agent Merge, multi-root sessions and new agent-session organization extend VS Code's AI tooling beyond code generation and into pull request completion, orchestration and review.