AI coding agents have become one of the fastest-growing categories in enterprise software. In the span of just a few years, these development tools have evolved from simple autocomplete assistants into autonomous systems capable of taking over the complete software development cycle, all via natural language prompts.
As vibe-coding takes off, tools from startups like Cursor and Anthropics Claude Code have quickly reached multibilliondollar revenue run rates. Cursor reportedly crossed $1 billion in annual recurring revenue (ARR) in 2025 and has since approached $2 billion in Q1 of 2026. Anthropics Claude Code has scaled even faster, reaching an estimated $2.5 billion annualized run rate within its first year, making it one of the fastestgrowing products in the category that accounts for a large share of Anthropics $14 billion ARR.
Yet inside large enterprises, writing code is rarely the hardest part of the job. Data scientists, engineers, and analysts spend much of their time maintaining and augmenting pipelines rather than building new ones. The real bottleneck in enterprise AI, therefore, is not software development itself, but operating complex data systems in production.
Databricks CEO and co-founder Ali Ghodsi believes that the gap represents the next frontier for AI automation. In his view, the next generation of AI agents wont just write software, but operate the data systems that modern businesses depend on.
That strategic bet is behind Genie Code, a system of autonomous AI agents unveiled today, designed for data engineering, data science, and analytics operations. The system extends the companys existing Genie platform ecosystem, which allows knowledge workers to ask questions about enterprise data in natural language. (More than 20,000 organizations already used Databrickss data management and analytics tools; the companys ARR surpassed $5.4 billion annual revenue in February.)
Instead of functioning merely as a coding assistant or helping generate code faster, these agents actually understand the structure of the data and existing data problems, Ali Ghodsi says. It can automatically set up pipelines, analyze why something is failing, and understand issues like when a dataset schema changes or when permissions are modified.
For instance, Genie Code can help determine how a dataset should be prepared for modelingrandomizing the data, separating part of it into a test set, or training a model on the remaining portion. After training, the system can aid in evaluating the results using metrics such as F1 scores or the area under the curve, and then analyzing them to determine whether the model is performing well or requires improvement.It can suggest trying different approachesmaybe retraining the model or generating plots and graphs to visualize performance, and uncover reasoning about what changes might improve the results, Ghodsi explains. Its not about just generating random code snippets, but understanding the entire structure of the data problem and working through the modeling workflow the same way a data scientist or engineer would.
Databricks and Enterprise Context
A major reason many AI coding agents struggle in enterprise data environments is context. Most developer tools train primarily on public code repositories and general programming examples. Enterprise data systems, however, add another layer of complexity. Data carries business semantics, governance rules, and access policies that determine how information can be used. Without that context, an AI agent may generate technically correct code that fails once deployed in production.
Genie Code attempts to address that problem by integrating directly with Unity Catalog, Databricks governance framework for enterprise data. This integration allows the system to understand data lineage, access permissions, and organizational policies across an enterprises entire data estate.
Maintaining pipelines and making sure they are reliable and always running is a big part of a data engineers job, and this is where Genie Code can augment them significantly, Ghodsi says. It can monitor systems continuously and respond immediately when something breaks, even in the middle of the night, analyzing complex traces and diagnosing what happened so that the pipeline can be fixed and kept running reliably.
The architecture relies on a multi-agent architecture powered by multiple AI models. Ghodsi explains that the system combines LLMs from providers including Anthropic, OpenAI, and Google, alongside smaller open-source models optimized for specific tasks. There are many things inside a workflow where you dont need a huge modelyou just need something fast that can perform a very specific operation reliably.
The larger models provide the reasoning capabilities necessary for complex problem-solving and planning. Smaller open-source models are trained to handle more routine operations quickly and efficiently. Moreover, the architecture is built around multiple collaborating agents rather than a single monolithic AI system. Each agent specializes in particular functions, such as diagnosing pipeline failures or analyzing data patterns. These agents share context, memory, and skills, allowing them to coordinate their actions and execute complex workflows across the data stack.
Databricks describes this approach as agentic data work. Rather than prompting an AI assistant for small pieces of code, users can delegate entire objectives to the system.
Another challenge with autonomous AI systems is maintaining reliable performance in production environments over time, as agents often encounter unfamiliar scenarios that degrade performance. To address that issue, Databricks has acquired Quotient AI, a startup specializing in evaluation and reinforcement learning for AI agents. The companys technology helps evaluate agent behavior, continuously measuring output quality and detecting regressions before they cause production failures. Quotient AIs founders previously worked on improving the quality of GitHub Copilot, giving them deep expertise in evaluating AI coding systems.
Vibe-coding for data systems
The rise of vibe-coding has created a new battleground for agentic AI-powered coding tools and reshaped the competitive landscape in software infrastructure. Databricks is approaching the market from a different direction. Ghodsi says the AI coding market and the enterprise data automation market are evolving in parallel but distinct directions.
While tools like Cursor and Anthropics coding agents are reshaping how developers write software, Databricks is focused on transforming how companies manage and operate their data systems. Even though our product name includes code, what it really focuses on is data work, Ghodsi says.
Genie Code targets the workflows that occur after data enters an organizations platform. By focusing on the data layer, the company aims to address problems that general-purpose coding assistants are not designed to solve. The other tools in the market help software engineers write application code, which is great, says Ghodsi, But for us the end goal is the data: transforming data reliably, and helping organizations work with their data.
Several organizations, including SiriusXM and Repsol, have already begun experimenting with the technology. SiriusXM uses Genie Code to help build and maintain internal data products, generate SQL queries, nd debug pipelines. According to Ghodsi, the company has reported around 20% productivity improvements in data engineering tasks. Genie Code assists engineers in creating data products with defined service-level agreements and reliability guarantees.
Likewise, multinational energy and petrochemical company Repsol is using the technology to accelerate forecasting and production workflows. Instead of manually connecting notebooks, pipelines, and models across different systems, engineers can rely on Genie Code to orchestrate these processes automatically. Ghodsi added that thousands of other customers are already experimenting with the technology, although many deployments are still in early stages.
The Future of Human Engineering
Ghodsi does not expect autonomous agents to replace human engineers. Instead, engineers may spend less time writing code and more time designing architectures, supervising automated systems, and ensuring that AI-driven workflows operate reliably.
The cost of automation is going down and the tools are becoming easier to use, so naturally the demand for automation increases. If you look at some of the numbers already, a huge percentage of activity on machines is actually agents operating in the background, he says.
According to the companys recently released State of AI Agents report, AI agents now create 80% of databases and 97% of test and development environments on the Databricks platform. Just two years ago, agents barely registered in database activity, with human developers handling nearly all of that work.
I wouldnt be surprised if that number goes from something like 80% to 99% in a short period of time. But that doesnt mean humans disappear from the process, Ghodsi explains. You also have to think about legal responsibility and quality guarantees. Those are areas where you still need a human in the loop.
Canva’s new AI tool, launching today, is going to save time, money, and headaches for so many people. Called Magic Layers, it turns any flat bitmap image into a fully editable Canva project, extracting text, objects, and components into individual layers.This tool marks a fundamental shift in how we handle digital assets. Until now, a rendered image was basically a locked vault of pixels. If you wanted to change a typo or swap a background, you had four options: 1) Hunt down the original project file, 2) painstakingly change it in Photoshop, 3) accept a generative AI patch job, or 4) close the laptop and escape to live a real life somewhere by a nice beach. Magic Layers shatters the vault. By reverse-engineering a flat picture into its constituent parts, Canva cofounder and Chief Product Officer Cameron Adams tells me, Magic Layers empowers users to resurrect and tweak any image they have on their hard drive.[Image: Canva]Canva uses many models from OpenAI, Anthropic, and other developers, but the secret sauce behind this new layering capability is its proprietary AI design model, which the company unveiled last October. Think of it not just as a random design and image generator, but as a model that understands the elements of design. It looks at a picture and sees its skeletal structuredistinguishing the foreground subjects from the background scenery, and recognizing typography as actual text rather than just colored shapes. When you feed it an image, whether it was spat out by an AI prompt or dragged from an old folder, it dissects those elements perfectly. The new Canva multilayer tool is the implementation of those abilities.Most AI outputs are fixed, really flat things, and they’re not easy to edit. You either have to, like, live with an 80% solution or you have to spend time reprompting, trying to get that little bit of the image that you wanted to get fixed,” Adams says. But now, he adds, the model identifies everything in the frame and converts it into native Canva objects.So text isn’t just a cutout anymore. It becomes a live, editable text box. You can correct spelling errors, swap the font, adjust the size, or even translate the copy for international markets. The same goes for visual objects. Once separated, elements like a product bottle or a butterfly become completely independent actors on the canvas. You can move them, resize them, change their color, or banish them from the composition entirely without leaving a gaping hole behind, Adams explains.And since these extracted layers are treated exactly like standard Canva design elements, you can apply all of the platform’s existing tools to them, including upscaling or generative tweaks like Magic Edit. That’s the beauty of it, that it’s now a proper Canva design. So you can change any of those elements in any way,” Adams says. Because Canva operates in the cloud, this newly resurrected file is immediately ready for multiplayer collaboration. You and your team can jump into the project simultaneously and start moving things around. [Image: Canva]Its getting better all the timeThere is an interesting parallel here with Adobes recent launch of a new AI assistant for its web and mobile Photoshop apps. Both companies are trying to fix the fundamental flaw of current generative AI models like Google’s Nano Banana.When you ask a standard AI to remove a single item from a picture, the machine recalculates the whole picture from scratch, inevitably introducing random errors or hallucinations. Adobe tackles this problem by allowing users to point at or draw around an object. The AI then places these modifications on independent, clear overlays suspended above the base image, preserving the underlying raw pixels flawlessly. While Adobe’s method builds new, highly controlled editsincluding texton top of an existing foundation to guarantee precision, Canva’s Magic Layers takes the opposite route: It dismantles the foundation itself, breaking the flat image apart into discrete, fully interactive components.While these tools from both companies do, indeed, appear to be magical, to me they feel like features that are not going to stick around for too long. Theyre more like patches that solve generative AIs current problems with output uncertainty.Once engines like Nano Banana or Seedream can nail down every pixel, every text and typography, every single human, animal, tree, pair of jeans, or shampoo bottle everand it will happenwe will no longer be worrying about things being in layers. Objects, type, and components will simply exist in the reality of the image; the models will understand them just like humans do, allowing users to change anything they want instantly, and with precision. Everything will be liquid for you to touch and change. Software will follow your exacting and most complicated whims with perfection. But for now, Magic Layers is going to solve a lot of problems for a lot of people and companies all around the world.
Each year, some of Americas greatest artists, thinkers, and business leaders have a chance to come together at SXSW in the spirit of creativity, innovation, and future-building. And with everything currently happening in technology and the workforce, this years gathering feels particularly timely.
Of course, questions around AI will take center stage and remain our primary cultural fixation: How long until the next incredible breakthrough? Should Americans be fearful about an impending AI apocalypse or hopeful about the prospect of unlimited productivity gains?
These topics are all valid, urgent, and deeply worthwhile to explore, but I also believe the most important workforce story unfolding in the U.S. today is less about what AI will do next, and more about what everyday Americans are doing right now in response to and in preparation for AIs growing impacts.
If technological advancement is going to keep accelerating faster than our institutions can or are willing to adapt, the fact that workers have already begun adapting on their own in real time is a story of deep-rooted resilience within our culture and communities. It is also a story that seems to be signaling a pragmatic and optimistic reimagining of the American Dream.
WORKFORCE DISRUPTION IS WELL UNDERWAY
The speed of AI advancement is likely to continue to be astonishing. Although we can neither predict nor control the pace of innovation, we can acknowledge that AI is no longer a hypothetical but an economic force reshaping job security, hiring, and career planning.
We also need to understand that while AI adoption has added pressure, workforce fragility in the U.S. was deepening long before generative models like ChatGPT entered the picture.
Education costs have been compounding at an unhealthy rate in America for nearly half a century, with rising tuition costs significantly outpacing inflation since the 1980s. Meanwhile, the countrys student debt crisis also continues deepening, with total student loan debt in the U.S. exceeding $1.7 trillion in 2024, all while broader confidence in traditional education and career pathways has been gradually eroding.
AI isnt causing workforce uncertainty but merely adding weight on top of existing cracks in the system. To focus solely on predicting the pace and extent of AI-driven job loss misses the real story: U.S. workers are already adapting, and its a process involving a bold reimagining of American values and stability.
AMERICANS ARE CHOOSING DURABILITY
Despite so much uncertainty, Americans dont appear to be giving in to fear as much as theyre leaning into resilience and practical decision-making. There are some strong cultural signals indicating a radical shift in the U.S. workforces strategic mindset, particularly in evolving views around traditional education and career pathways in this AI age.
More specifically, a new survey of American workers we conducted at the Business For Good Foundation via the Harris Poll revealed a clear and widespread departure from most conventional ways of thinking about professional and economic fulfillment. For example, 75% of Americans shared that their views of a good job does not look the same now as five years ago, while 80% agreed more people are choosing trade training over four-year degree programs.
Similarly, more than 78% said they believe long-lasting social and cultural stigmas around blue-collar work are beginning to dissipate in the U.S., with 76% saying they believe trade jobs are less likely to be replaced by AI.
Rather than fearing widespread job loss and sustained unemployment, Americans are envisioning a future workforce defined by durability, where the workforces economic value is concentrated less in white-collar sectors and more by the durable, hands-on skills that have always played an indispensable role. It suggests an overall mood of pragmatic optimism, with Americans appearing to adjust to AI adoption much faster than our political and educational systems.
GET AHEAD OF CHANGE
While everyday Americans seem eager to get ahead of AIs inevitable changes, this likely wont happen at scale without the appropriate support from organizations and U.S. business leaders. Recognizing this heightened need for more hands-on programs to increase access to skilled trade training, we at the Business for Good Foundation committed $100,000 to advancing workforce development in the first half of 2026.
Of course, this will also require strategy and coordination, grounded in shared recognition that this shift away from traditional white-collar pathways is not an error but a process of economic regeneration. The growing emphasis on hands-on trades is not nostalgia, but necessary to strengthen the U.S. innovation infrastructure.
Skilled work continues to underpin all non-negotiable aspects of American society, including access to housing and healthcare. At the same time, U.S. business owners are grappling with critical, pre-existing skilled labor shortages, meaning theyll increasingly need to depend on talent pipelines beyond traditional degree models.
One recent example of what weve done at the Business for Good Foundation is a New York Capital Region pilot. As part of our commitment to workforce development, the foundation awarded a $25,000 grant to the Social Enterprise and Training (SEAT) Center to expand trade skills programming in the region and help bridge the gap between untapped talent and industry demand.
Ive seen firsthand that simple, practical investments like in the SEAT Centerthose that better align workforce pathways with employer needs and expand access to education and career opportunities for motivated talent in underserved communitiescan go a long way toward creating a real and sustainable path to upward economic mobility. Im encouraging leaders across the country to take similar action, at any scale.
However, such a model will largely remain limited without other like-minded business leaders and philanthropists willing to build on and replicate it at scale, and who are prepared to fully embrace a new American dream defined less by credentials and more by individual capabilities, determination, and human resilience. While this kind of change certainly wont happen overnight, I hope that those of us who attend SXSW this week might begin aligning our business priorities with the unique spirit of this event, working together to intentionally build a brighter, more prosperous, and innovative future for the U.S. workforce.
Ed Mitzen is cofounder of Business for Good Foundation.
Apples iOS 26 for iPhone got off to a rough start when it was finally released to the public in September of last year.
Its new Liquid Glass design language remained unpolished in many areas, and the operating system harbored a fair amount of bugs. But since iOS 26.0 debuted, Apple has released three major updates for it, further polishing the interface and adding new features.
And soon, Apple will update iOS 26 once again with the release of iOS 26.4. Its a release that is set to not just eliminate bugs and enhance the details of Liquid Glass, but is also set to add some significant new features to your iPhone.
Heres whats coming, and when you can get iOS 26.4.
What new features are coming to iOS 26.4?
Apple has been beta testing iOS 26.4 since last month. Originally, the software update was rumored to include the companys revamped Siri, powered by Googles Gemini LLM. However, Siris AI revamp has been absent from all iOS 26.4 betas to date, so it looks like a truly useful Apple digital assistant is still a ways away.
But that doesnt mean iOS 26.4 doesnt have any new features. Quite the contrary. Besides your normal user interface polishes and bug fixes, iOS 26.4 is set to include some major upgrades to its media apps, notes 9to5Mac. Those upgrades include:
AI-powered music playlist creation: iOS 26.4 will add a feature to the Music app called Playlist Playground. The feature allows you to generate music playlists from natural-language text descriptions. So you could instruct the Playlist Playground feature to make a playlist of 80s rock ballads under five minutes long, and the Music app will generate a playlist based on your prompt.
Podcasts app video overhaul: Apples Podcasts app has supported video podcasts for some time. But in iOS 26.4, its video capabilities are getting a major upgrade. Now you can quickly switch between the audio and video versions of a podcast. This feature will be great for those times when you are watching a video podcast, but then suddenly need to be on the movesoon youll be able to easily switch to the audio version of the podcast, ensuring you can still enjoy it when your eyes are needed on other things.
Redesigned album and playlist interface: Also in the Music app, Apple has redesigned the look of the interface that you see when displaying playlists or albums in full screen. In iOS 26.4, the Music app will now tint the entire screen based on the album art color scheme, giving each playlist and album its own unique look.
And thats not all: iOS 26.4 will add numerous small refinements and additions across the operating system, including new emojis, new Ambient Music widgets for your Home Screen, automatic activation of Stolen Device Protection, and more.
iOS 26.4 beta: Download it now
While Apple hasnt released iOS 26.4 to the general public yet, it has released four betas of the software to developers and public beta testers. And if you are in any of those two groups, you can download the latest beta of iOS 26.4 onto your iPhone today.
To download the developer beta, youll need to be a member of the Apple Developer Program.
If youre not a developer, but still want to try out the new software early, you can join the Apple Beta Software Program for free and get access to public betasincluding the iOS 26.4 betatoday.
Of course, the usual warning applies: betas are buggy, and in rare cases, they can cause data loss or otherwise harm your phone. So always proceed with caution if you decide to download a beta.
iOS 26.4 final release: Download it later this month
If a beta isnt your thing, youll have to wait until Apple releases the final version of iOS 26.4 to the public. Thankfully, you probably wont have to wait too much longer.
Apple generally has a 5-6 beta development cycle for iOS point upgrades like iOS 26.4. Apple released the first iOS 26.4 beta in mid-February, which means the final public version of the beta is highly likely to be released between mid-March and the end of the month.
Once Apple releases the final version of the software, youll be able to download iOS 26.4 right to your iPhone using the devices Software Update feature in the Settings app.
OpenAI confirmed on March 6 that it is delaying the rollout of adult mode in ChatGPT, a feature that would give verified adults access to less-restricted content. The company first announced plans to begin age-gating users last year but has now pushed back the launch twice. Segregating adult users from minors could help in some of OpenAIs legal and revenue challenges, but nailing the technology may not be easy.
Adult mode had been expected this quarter and still is, just later than originally planned.
OpenAI referred Fast Company to a comment it gave to Alex Heaths Sources newsletter saying it was pausing the feature to focus on improvements to ChatGPT, including gains in intelligence, personality improvements, personalization, and making the experience more proactive. (It also told Axios it needs more time. We still believe in the principle of treating adults like adults, but getting the experience right will take more time, the company said.)
OpenAI first hinted at the feature last October in an X post from CEO Sam Altman responding to questions about ChatGPTs safety for underage users. As we roll out age-gating more fully and as part of our treat adult users like adults principle, we will allow even more, like erotica for verified adults, Altman wrote.
Adult mode depends on OpenAIs new age-prediction and verification system, a homegrown AI model that estimates a users age based on prompts and media generated with tools like Sora. OpenAI said in January it had begun rolling out the technology globally within ChatGPT. When the system detects a user may be underage, it restricts things like violent content and romantic role-play.
The company also uses a third-party verification platform called Persona, which allows users to confirm their age if the AI system places them in the wrong category.
But accurate age prediction and verification isnt easy, and using it to age-gate chatbot users is a relatively new idea. How, for instance, can an AI model distinguish between a 16-year-old in high school and a 19-year-old in college when the two talk to ChatGPT about similar things?
“You can imagine . . . if you input anything that looks like a homework question, ChatGPT flags you as a minor, and then you’re automatically in the minor bucket, says Alissa Cooper, executive director of the Knight-Georgetown Institute, a tech policy group. That would pretty seriously constrain the service for college students or people who might just happen to have asked a question that looks like a homework question.”
Naturally, some younger users will try to trick the age-prediction model into thinking theyre adults. There’s not really a way to prevent circumvention regardless of the architecture or the system design, Cooper says. So there’s this balance between locking things down to try to prevent circumvention, and allowing a full-featured experience for users who are age appropriate for whatever that experience is meant to be.
One of Coopers main concerns is that the world outside OpenAI wont know how well the system is performing. And, so far at least, OpenAI isnt sharing much information about its system.
I think it’s correct to be skeptical, Cooper says, adding that she believes companies should provide enough transparency about how their age-verification systems are tested so that independent experts can evaluate whether they actually work and examine the data used to estimate the ages of potentially hundreds of millions of users.
OpenAI was put on notice last year when it was twice sued by people claiming that earlier versions of ChatGPT had led an adolescent loved one toward suicide. This set the stage for the company first applying more guardrails to its models for all users, then attempting to cordon off a safe experience for younger users. With younger users safely segregated, OpenAI could loosen or remove some content restrictions for adult users, the thinking goes.
That adult mode could become a real selling point, and right now ChatGPT could use one. ChatGPT, which has 800 million weekly active users, once faced little real competition among AI chatbots. ChatGPT still has the most users, but competition has heated up with Googles improvement of its Gemini chatbot, and with Anthropics Claude gaining more mainstream name recognition.
Not only is OpenAI under pressure to fend off those rivals, but its also under pressure to increase revenue from the chatbot to help offset the massive expenditures it plans to make in new data centers over the next five years.
Signing up millions of new adult users to that experience would not only increase OpenAIs subscription revenues, but it could mean millions more highly engaged eyeballs to look at advertisements (the company said in January itll soon start showing ads to some of its U.S. users).
So I think it’s segmentation of the user base in multiple directions, says Cooper. Its keeping minors away from experiences that nobody wants them to have, but also being able to offer adults experiences that are truly adult-oriented that some adults want to have.
AI disruption and geopolitical upheaval are forcing business leaders to make high-stakes decisionsfast. Accenture CEO Julie Sweet shares what she’s hearing from her 9,000 clients, and the hard-won advice she’s giving them. Sweet reveals why AI proficiency is now a requirement for promotion at Accenture, why she’s doubling down on entry-level hiring amid the automation wave, and she unpacks the hidden power of “leader-led learning.
This is an abridged transcript of an interview from Rapid Response, hosted by the former Fast Company editor-in-chief Robert Safian. From the team behind the Masters of Scale podcast, Rapid Response features candid conversations with todays top business leaders navigating real-time challenges. Subscribe to Rapid Response wherever you get your podcasts to ensure you never miss an episode.
Accenture works across 120 countries, 9,000 clientsyou’re in every industry. You have this unique visibility into how organizations and leaders are navigating what is really a chaotic, fast-moving environment. Are there questions that you’re hearing particularly often right now?
Well, Bob, it’s interesting if I just start with Iran, because I’m getting a lot of questions, particularly in Europe, where if you think about a potential energy crunch, it’s expected to hit harder in Europe than, say, the impact on the U.S. Everyone believes that this environment, where energy is a risk, is just their new norm. And actually, there’s more optimism because if you compare this to 2022, when the war in Ukraine started, Europe is in a much better position from a resilience perspective.
And it’s a theme that we’ve been seeing for quite some time. I got the same questions even a couple of months ago when we had this whole issue around tariffs and imposing them, which is that CEOs are really just expecting the unexpected. It’s being built in, and that’s why resilience is such a big theme. There are also big questions continuing on AI, et cetera, but I wanted to address the latest, which is the impact of the Iran war.
When I talk to CEOs right now, there’s this sense that some of them seem almost frozen. They’re waiting for clarity. And I know you’ve encouraged the opposite: Don’t take cover; take chances. How do you know when to act and when to wait?
The reality is, as a CEO you can’t bake anything into your plan, simply because so much is unknown. And that’s where transparency really matters, being able to say, “Here’s what we know. Here’s what we don’t know.” And then we’re making our action plan with those things in mind.
So when you think about when you know to act or not act, in my view you’re always acting. It’s intentional decisions. It’s an action to say, “Because I don’t know this, I can’t alter my plans.” One of the biggest risks right now is even less about the impact on the economy from what’s going on in the Strait of Hormuz and energy. The bigger risk that many companies are talking about is, do we have a cyberattack or an attack on critical infrastructure that spins out of that?
And for that, there are actions you can take, because we’re helping clients look at their cyber resilience. This has already been a big growth area for us because AI itself has increased the attack surface. When you think about what the risks are, you can’t really tell what’s going to happen on the energy front. But knowing what you do know, what you don’t know, and whether there are actions that can be takenand then actingmatters.
Alongside all this geopolitical activity, there are also the technological shifts that we’re experiencing. You made a bold move last year where you merged a bunch of divisions into a single unit that you call Reinvention Services around AI. You’re coming off your biggest-ever quarter of revenue growth in new business. We see businesses that are not tech businesses that may not be getting the return on investment from their AI that they had expected or hoped for. Does it work in some places and not others? Are there industries and functions that are more fit for AI than others? Or is it more about culture and commitment?
Well, in some ways it’s going to be all of the above. So first of all, we have to remember the technology, while changing fast, is still really early. And there’s a lot of actual value that companies can get before advanced AI. So let me give you an example.
We work with a pharma company that takes drugs to market, and there’s a process that, once the drug is approved, involves lots and lots of regulatory work. So what you write to explain it to the physician takes a long time. We’ve said . . . that you could actually shorten that. Instead of it being months, it could be much faster if you changed the process to be standardized, if you kept your data in one place. None of that requires generative AI or advanced AI, but for most companies, they haven’t done that.
So when we worked with [this company], advanced AI was the catalyst because it enables you, if you have all the standardized processes, to actually create the content faster. But the first piece always could have been done and hadn’t been done. And so much of the work that we’re doing is actually work where companies are saying, “Okay, wait a minute. Before I spend the money on advanced AI, I should clean up my fragmented processes. I should standardize things. I should not have as many people in middle managementcompletely apart from agentsbecause why would I spend money to create an agent to replace a manager that I shouldn’t have in the first place?”
Is part of the hope that the motivation for not having done this is, “Oh, the agent or AI will come in and do this for me. I can skip a layer, save money.” And that’s the silver bullet. And if I’m hearing you right, you’re like, “Yeah, maybe not.”
Yeah. I think there was this view at the beginning where companies thought, because it’s so easy, are they just going to do it all? Do I just ask this model, and this model is going to tell me how to change my company? One hundred percent, that is not it. The models don’t know how to change a company. And if you spend money on your current structure just replacing parts of it, you’re at best going to get incremental value.
The real value is to actually reinvent everything you do. And that reinvention doesn’t start just with advanced AI, but with a lot of really basic lessons where companies haven’t had the will to fix things. I tell CEOs every day: In three years, you should be able to [answer], What did I use AI to make possible that was impossible before? Because if the only thing you’re doing is incrementally improving how you’re operating, you’re not going to get the biggest value. The biggest value is in the core operations of a companythings you’re going to be able to do with asset manageent in any industrial company, things you’re going to be able to do with the grid in utilities.
The tech isn’t completely there yet. It’s still error-prone. That’s why everybody’s tracking things like how long the tech works and the memory piece of it. So where you get value today is anything with customers, because those are short interactions. The tech has to continue to improve, and the strategy has to be, “I’m going to use this tech to do something I couldn’t do before.”
There’s this phrase, “AI-first.” You’ve described Accenture as an AI-first company. It’s something a lot of other players aspire to but struggle to implement. Is it hard to be AI-first? What does it mean?
First of all, it is hard. And the reason it’s hard is that it requires your leaders to understand what AI does. And this is so different from the digital era. Moving to the cloud and stuff, a lot of it was plumbing. So as a leader, you didn’t have to understand it because it was being handled by the tech folks.
To be AI-first, you have to say, What can AI actually do? So you have to understand things like, wait a minute, it has to have a certain amount of memory to be able to do something. You have to understand what it’s actually able to be accurate about. And then think about your business to say, “Where can I get a big enough return for using something at this cost?” It starts with leaders having to understand technology in a totally different way.
When ChatGPT first emerged in November 2022, the people who received the most training initially were my top 50 leaders, because I knew that if they didn’t understand the power, they would not be able to help us transform how we’re delivering our services and what our clients could use it for. So leader-led learning is a huge unlock. And then AI-first is asking yourself, Is this something that AI could do?
In todays world, the villain in our story isnt a person; its our desire for instant gratification. Explosive sales growth? We want it now. An dream angel investor? We want it now. A raise, a promotion, a spot at the top? We want it now.
Can you blame us? If we can binge-watch an entire season of a new show on Netflix in a weekend and order restaurant-ready food to our door in less than thirty minutes, that can set us up for unrealistic expectations about getting other things quickly, including in the workplace.
The need for speed leaves us rushing and impatientand it shows in the way we speak, too. Our conversations become transactional, our questions become shallow, and our communication prevents us from building trusted relationships with those around us. If you’re nodding your head, I invite you to consider three conversation-killers to avoid.
1. Conversation domination
Two words: talk time. If the amount of time youre talking is more than the person youre communicating with, youre dominatingand it can be detrimental.
If youre using the precious commodity of time to push your agenda, solution, or unsolicited viewpoint on somebody without solicitation, youre talking at them, not with them. Often, we dont even know were doing it. Plus, when we’re housing nervous energy, we can unknowingly engage in conversation domination as a way to soothe our internal discomfort.
The solution? Create a personal practice to ground yourself before every high-stakes conversation so you can experience more clarity, calm and presence. For example, if youre in a season of feeling time-poor, try the physiological sigh. A technique that was discovered in the 1930s to help us rapidly regain control from feelings of stress or anxiety, simply take two deep inhales through your nose and one long exhale through your mouth with pursed lips. Almost instantly, youll experience less tension and a sense of presence. Repeat it a few times if needed. This will help you catch yourself in the act or prevent conversation domination altogether.
2. Trying to be interesting instead of interested
Dale Carnegie once said something along the lines of, To be interesting, be interested. Heres how I see it: in any given conversation, your job isnt to make yourself look significant; its to make the person opposite you feel significant. But how do you do it without feeling contrived?
Consider conscious questions, which as I define it, are questions that are grounded in positive intentionality. For example, you could walk past your colleague and say, Hey Mark, how are you? Or you could say, Hey Mark, you mentioned the other day that you were stressed because you had to take care of your sick son while preparing for that big keynote. Hows he doing? How was the speech?
Do you see the difference? The former lacks depth. The latter is a meaningful question that exhibits intention. Do this right, and youll show others how youre interested in what theyre emotionally invested in.
3. Being attached to an outcome
Whether youre in a job interview, a sales call, or a meeting with leadership, the stakes can be high. But if you enter any of these conversations attached to a specific result, youre likely to act inauthentically. Your body language, tone, rate of speech, energy, and more will unconsciously map to your need for an outcome (and often rushing to a specific timeline). That can undermine your communication.
Say youve set a professional goal of landing a promotion within the next twelve months. Fast-forward eleven months, and there you are, sitting in a meeting with leadership, discussing a potential promotion. Instead of asking intentional questions, deeply listening, and being truly present, the timeline in your mind has you feeling pressured, impatient and or reactiveand others can tell. You sabotage your own success.
Heres an alternate approach: Ask yourself, If I were overflowing with abundance in every area of my life, how would I behave in this moment? Once you remove your attachment to an outcome, you create an openness to receiving what is truly meant for you, even if its not in the time or path you desire.
What this means for leaders
If an organization wants to build a high-trust culture, increase employee engagement, and create a sense of belonging for their people, it begins with leadership learning how to have conscious conversations. The key lies in embodying the behavior you want others to exhibit.
Psychologist Albert Banduras social learning theory (SLT) suggests that people learn new behaviors by observing and imitating others. Simply put: when we observe the consequences of other peoples behavior, were more likely to imitate the actions that are positively rewarded and avoid those that are punished. In turn, this leads to an acquisition of knowledge, attitudes, and beliefs.
If leadership embodies conscious, non-transactional communication, and rewards others for following suit, this will create change at scale, and a high-trust culture will diffuse as if through osmosis. The byproduct? Long-term success thats built on the right foundation.
In a remote-first world, we can solve problems, build teams, and maintain relationships from behind a screen. But thanks to those very screens, human connection and communication matter more than ever. The leaders who will stand out are those who prioritize them.
Adapted from Relationship Currency by Ravi Rajani. All rights reserved.
We are cooked.
That’s the sentence I see with every AI-generated Instagram, TikTok, or YouTube short made with Seedance 2.0. And yes, we are. The walls of reality have finally vanished, sucked in by a black hole of Nvidia chips. So I’m going to Nancy Reagan the hell out of everyone and demand a global public service announcement like that old Just Say No to drugs campaign, which was everywhere when I was growing up.
We need Mr. T back to make young and old fools listen up, because the companies printing money with their generative video tech are doing zilch to fix the planetary problem they have created.
The message? Everyone should stop believing everything that moves online. Or at least question it all with a critical mind. All the time.
It will be hard. Probably impossible. The instant satisfaction of buying into whatever candy social media throws at us, algorithmically tuned to support our preconceived ideas, is too much to resist. We want to believe because dopamine is so yummy. And the digital overlords of Silicon Valley and Beijing know it. Thats why they have officially trampled our already fragile grasp on the truth with the release of models capable of manufacturing clips that are indistinguishable from physical life.
AI models like ByteDance’s Seedance 2.0 can wolf down up to a dozen reference filesimages, audio tracks, and camera movement samplesto flawlessly synthesize an alternate reality with no uncanny valley. And it costs only pennies do so.
We have effectively handed the keys to the multiverse to any basement-dwelling sociopath with a Wi-Fi connection. Tal Hagin, an information warfare analyst, told Euronews exactly where we stand: We are no longer at the stage where it’s six months away. We are already there: unable to identify what’s AI and what’s not.
The same computer industry that has destroyed the space-time fabric has failed to deliver its Content Authenticity Initiative, which promised a way to certify and label truly real videos. Imagine that. So someone needs to educate people to doubt everything they see online.
If you think Im exaggerating the immediate danger, just look at the circus of Nicolás Maduros capture by U.S. Special Forces in January. There was no Seedance 2.0 then (less than two months ago!), but social media was instantly paralyzed by a flood of highly realistic, completely believable AI-generated images of the ousted Venezuelan leader.
Across X, TikTok, and Instagram, synthetic media of Maduro in custody or crowds of Venezuelans celebrating racked up millions of views in mere hours. Millions of peopleincluding the usual politicians and tech billionaires whose thumbs are perpetually superglued to the retweet buttonswallowed the digital slop whole.
Primeras imágenes de Nicolás Maduro capturado. pic.twitter.com/d8RjDNC3zm— SheIby (@TommyShelby_30) January 3, 2026
Hagin noted that the moment an information vacuum opened regarding Maduro’s capture, individuals started uploading AI-generated images of Maduro in custody of the U.S. Special Forces in order to fill that gap.
The most worrying stuff is not those big news moments, which will get fact-checked promptly. Its the little things, the daily stuff that will have greater impact on our psyches. The local news, the scams, the bullying in school, the gossip about that neighbor everyone hates, the teacher, the office enemy, the ex-partner . . .
When reality breaks, replaced by a manufactured one, everyone will suffer.
So I’m calling for the Mother of All PSAs right now. We cannot sit around waiting for the tech industry to self-regulate, because history proves its leaders possess the moral compass of a weather vane.
We need a massive, impossible-to-ignore, flashing-red-light educational campaign pounded into the retinas of every smartphone user on Earth. We need to grab the public by the lapels and shake them until they finally understand that their own eyes and ears are now compromised enemy combatants.
So lets do that. Let’s not assume that people will eventually get it because millions of lives and minds are at stake. For the next year or so, let’s launch a worldwide education campaign where every commercial break, every YouTube pre-roll, and every TikTok swipe features a brutal, relentless reminder that objective reality is officially a relic of the past.
Everyone must build up and wear psychological armor like we are living in an MMORPG from hell. This needs to be the 21st-century equivalent of Stop, Drop, and Roll, except instead of being physically on fire, your perception of truth is being incinerated by a server farm in Guangdong. We have to normalize radical skepticism before its too late.
But since nobody is going to do that, just remember, kids: Don’t believe everything you see. Love your mama. And don’t do drugs. Or do drugs because realityis not real. Who the hell cares anymore?
Networking as a solopreneur can feel impossible. LinkedIn is full of the sort of hustle-culture aficionados who think yoga at 4 a.m. is something to brag about and who want you to buy their online course. Joining a networking referral group often costs money and can require a big time commitment without a guarantee of new leads. Asking friends and family to make referrals for you gives you flashbacks to that one summer in college when you got roped into selling Cutco knives.
But solo businesses are already nontraditional, so you might as well embrace quirky networking opportunities. Some of my best freelancing leads have come from Tumblr, carpooling, and on one memorable occasion, the ladies room at a Nick Cave concert.
If youre struggling with how to grow your network as a solopreneur, here are some unexpected strategies you can use.
Invite yourself in
When consultant Garima Verma wanted to break into the entertainment industry as a student at UCLA, she found that going to networking events and applying to every opportunity got her nowhere. So she decided she needed to get herself into the same room with the people she wanted to work for.
There was an event that NBCUniversal was sponsoring, Verma says. I wasnt invited to it and I had nothing to do with it, but I volunteered to help set up and clean up the chairs.
That meant she was there with the representatives at the end of the event and could get some one-on-one time with them. I was cleaning up and ended up talking toa little bit corneringa couple of reps, she says. Its how I got my first job in entertainment.
Verma has carried that same energy throughout her career. In 2020, she realized she wanted to do more in the world and got really deep into the volunteer infrastructure of the Biden-Harris campaign. I just DM’d a million people on Twitter and told them to talk to me and give me an interview for a job, she says. That’s how I got my first job in politics.
These days, Verma works for herself as a strategic advisor and consultant, but she continues to open her own doors. I don’t tend to get invited in the same way others might, and at a certain point I decided I’m going to invite myself in.
Ditch the elevator pitch
Author and speaker Jason Vitug talks about the networking anxiety that occurs in business environments. When youre expected to schmooze and impress other people, rather than simply connect, it puts too much pressure on every conversation. You might as well imagine Alec Baldwin telling you that coffee is for closers.
Thats why Vitug was able to feel comfortable chatting with someone at what could have been a disastrous book signing. No one showed up, and his new contact wandered over to ask why Vitug was sitting there. The two men enjoyed a spirited conversation that landed Vitug a speaking gig.
The bookstore environment allowed for a casual conversation, Vitug says. So my advice is to always be open to a conversation because theres a good chance if youre in the same place you have something in common.
While Vitug certainly offered his new patron some form of his elevator pitch during their long talk, he didnt lead with it. Instead, he was open to making a real and friendly connection to someone who was curious about him.
Immerse yourself in community
Charlotte Baker provides full-service payroll for small businesses in Jacksonville, Florida, with her new solo business, Easy Pay. When she was getting the solo enterprise off the ground two years ago, she heard about a community of local businesswomen that she wanted to join.
Women Business Owners of North Florida is an independent group, Baker says. Its not like a franchise or a paid networking group. Its a group of about 150 women whove all joined the organization to support and encourage each other.
Unlike the traditional group networking model, Bakers community does not expect members to bring referrals each week. Instead, the group offers weekly get-togethers that foster personal and supportive relationshipswhich Baker has found to be invaluable both emotionally and professionally.
Most of the women in the group don’t need my services, Baker says. But Ive built close friendships with these women, which has made my life a hundred times better as a business owner. And at least 60% of my revenue I can trace directly to recommendations from that group.
Becoming part of a supportive community makes networking much less onerous, since your friends will recommend your business, just as you will recommend theirs.
When networking looks like fun
Networking as a solopreneur only feels agonizing if you assume it has to follow the corporate rulebook. Theres no reason you have to post performative dreck on LinkedIn, show up at networking events in powersuits, or stumble through memorized lines about your solo business to expand your network.
Start by inviting yourself in. Whether you find a way to volunteer for an event that will put you proximity with someone youd like to talk to, or you keep knocking on doors (or sending DMs or emails) until you find someone willing to chat, remember that you can be friendly and persistentas long as youre willing to graciously take no for an answer.
Then ditch your elevator pitch. Remember that youre just a person who can have casual conversations with other people. Leading with curiosity and interest rather than a business agenda is more likely to end with a new contact. Its also much less nerve-wracking than self-consciously trying to network.
Finally, immerse yourself in your community. A large and supportive community will help do your networking for you, since people who care about you and believe in your business will naturally recommend you when they meet others who need your services.
Doing all of that makes networking something you can enjoy rather than something you have to suffer through.
Earlier this year, I had coffee with the chief investment officer of a large public pension fund. His fund doesnt invest directly into venture (they have a fund of funds position instead), so my new CIO friend doesnt usually get pitched directly by VC funds. He doesnt spend a ton of time in tech circles either.
When he does dip his toe in VC waters, he gets culture shock.
I have trouble understanding VCs, he said. (Im paraphrasing.)
By his estimation, people in traditional finance are easier to read. Their goal is to maximize returnsand the progress toward this goal is concrete, transparent, and measurable. Its really easy to understand what an asset managers motivations are when youre across the table from them in a professional capacity.
People in politics are also easier to read. Their goal is to build power and wield influence. So when you talk to them, you can assume thats what theyre looking for in the relationship.
Of course, both characterizations are limitingI know bankers who care about impact and at least one politician who cares about people (hes my cousin, so I can vouch). But as far as sweeping generalizations go, I can see where CIO is coming from.
In sharp contrast to financiers and politicians, VC investors are slippery creatures. CIOs have a hard time decoding our language. Venture capitalists are asset managers, but we talk like superheroes. We speak in hyperbole and aim, unironically, to change the world. We are incessantly crushing it, even though our portfolios are laughably unprofitable. We sit on boards but dress in jeans and sneakers. We are herd animals who claim to be contrarian.
Its hard for a CIO to judge how much of it is serious and how much of it is bullshit. And really, can you blame him?
We sound like this because of founders
I had a good laugh listening to that CIO, seeing this portrait of my industry from the eyes of one of its capital originators. But I do have a theory of where this language comes from, and why its mostly legit.
It starts with founders.
For most people, founding a companythe kind that scales massivelyis an irrational choice. Its extraordinarily difficult. You could be making way more money and working way fewer hours doing almost anything else. Chances are that youre going to fail, and youll have a pretty miserable time of it in the process. You have the odds of success of a lottery ticket, except that this particular lottery ticket costs 100% of your time, attention, and resources.
Nobody in their right mind would do this for the money. There simply has to be a greater purpose. And for founders, there usually is: a problem they are compelled to solve. A mission they feel called to achieve. A chip on the shoulder and something to prove. Sometimes, they simply cant imagine doing anything else with their lives.
Take it from an economist: These are all economically irrational reasons.
You literally cannot buy a founders time with stability and a high salary. Its why founders rarely sound like mercenaries or power-hoardersbecause theyre neither. They are motivated by something much greater. And to rational people like the CIO, it all sounds lofty, bordering on ridiculous.
Note, however, that this irrational exuberance makes for better, more resilient companies. It inspires angel investors and early employees, who forgo salary and stability for a dream. It keeps founding teams motivated for way longer than money alone does. Sometimes it even attracts customers and builds loyalty. Because a resonant mission takes you places that money alone cannot.
In other words: In our industry, irrationality is a feature, not a bug.
Venture is not a rational asset class
VC investing is also predictably irrational. VC funds are not capital conservation vehiclestheyre long-term illiquid, unpredictable, and alpha-seeking. There are thousands of other, safer ways you could be deploying your capital, so when you choose VC, you do it for the dream. To quote Recast Capital founder and managing partner Courtney Russell McCrae: “Nobody invests in venture to make median returnswere all aiming for the top, plain and simple.”
Thats what my CIO friend said, too. He said his company invests (a very tiny portion of its AUM) in venture because it is the only asset class that offers unlimited upside. Its the lottery ticket of finance.
Asset managers sell a product to limited partners (LPs). VCs sell a dream. The same dream that founders sell to us.
And that is why we all sound a little kooky.
Not all VCs are equal
Last year, I went viral for saying that megafunds are no longer venture capital funds. My argument is that theyre investing in consensus founders and consensus companiesnot in early-stage, high-risk, contrarian bets. Their largest deployments are into companies that are all but foretold to be winnersliterally too big, with too many giant powerful stakeholders, to fail. The bulk of their assets are being invested later and expected to generate faster and more predictable returns.
In finance, they call this type of risk “beta.” Its fundamentally different from the “alpha” risk you underwrite when you invest in day-one, early-stage, non-consensus founders.
These days, megafunds are making gobs of money on beta-seeking models. And it begs the question: Why do they still sound like VCs? Why do they want to hold on to the venture capital nomenclature, even when VC is a tiny proportion of their portfolio, just like CIOs? What do they lose if theyre called something else?
It occurs to me that these guys fundamentally dont want to be just bankers and stewards of capitalthey want to be visionaries. Certainly, theres a coolness factor, and the influence that comes with investing in the bleeding edge. But also, I bet you can measure the difference between banker and visionary by the size of their management fees.
For the record: I run a microfund, a fundamentally different vehicle and strategy than a megafund. I do not believe our funds should be analyzed togetherthey are fundamentally different assets, and warrant separate allocations, where you can compare like with like. If youre an LP, you are making bad decisions if you bucket all types of funds into a single giant VC bag. Youve been warned.
Boutique VC is an irrational choice, too
Speaking of irrational: Raising an early-stage microfund is an irrational choice, too. When you make all your money in carry, and very little in fees, youre betting completely on the upside, the dream. In the short term, you could be making way more money elsewhere.
Thats why I see the same motivation among emerging venture capital fundsor boutique VCs, as the megafunds prefer we call ourselvesthan I do in founders. Nobody chooses to do this for rational reasons. We do it for unlimited upside. We do it for mission or lov of the craft. We do it because the future of technology and the future of humanity are all being written by early-stage startups and scientists and inventors and R&D labs, and we want to have a say in it.
I personally do it because it is the purest incarnation of the American dreamthe idea that anyone can be the next founder to change the world, whether theyre consensus or not. This is what drives me. Its why I immigrated to America in the first place.
I know now what I sound like when I say this.
Maybe my pension fund friend is right to be confused. Maybe we do all sound like were full of shit sometimes.
But the reason we sound like thisthe reason we talk about doing good and having impact and changing the world and making a differenceis because some of us founders and VCs actually mean it.
And we wouldnt be doing this otherwise.
This story was originally published in Leslie Feinzaigs Venture with Leslie newsletter.
