This week, you can follow other learners, choose your own portfolio URL and put your learnlists in the order you want.

🤝 Follow learners and shape your portfolio

Portfolios were something you could look at but not keep up with. Here's what changed.

Follow learners from their portfolios

You can follow another learner from their portfolio with a Follow button, and see how many followers they have. Tap the count to open a list of who follows them. The count updates as people follow, and the badges in search match the Follow button.

Choose your own portfolio URL

You can pick the web address for your portfolio. New learners choose it with their photo when they set up their account, and anyone can change it in Edit portfolio. On short screens, that window couldn't scroll, so Save was out of reach. It scrolls now.

Reorder your learnlists

If you own a learnlist, you can drag its projects into the order you want people to work through them. The other projects slide aside to show where it'll land. Thank you to everyone who requested it!

✨ Smaller improvements

• The chat where you generate a project has a fullscreen toggle, so you've got room to read longer replies.

• The share icon on a portfolio shows "Copied" after you tap it, so you know the link is on your clipboard.

  

• Switching between topics in Explore slides smoothly instead of jumping.

  

• When you generate a project, it skips outline questions you've already answered in NextWork.md, the note in Settings where you tell NextWork about your background and goals.

🐛 Bug fixes

• On phones, tapping the email submit button twice during onboarding crashed the screen. It now handles the second tap.

• In exported PDFs, the fade over the top image covered the author line. The byline now stays readable.

• In lessons on Safari, folded code blocks wouldn't open. They expand when you tap them.

• Search rows had lost their progress rings, so you couldn't see how far along a project was. The rings show on every row again.

• When you searched for a project to add to a learnlist, projects you'd generated didn't appear. They show up in the results.

• The outline step of project generation sometimes failed if the AI returned extra details. It now carries on and gives you your outline.

• Opening the login window from a project's menu left the menu open behind it. The menu closes first.

🚋 In progress

• Reopen your past chats when you generate a project, so you can pick up where you left off.

• A work page you can edit freely, with a slash menu and templates that change its voice.

• A rebuilt settings page.

• Recording tools that capture your build as you work through a project.

🧪 New projects

• Route AI Requests with LiteLLM

  Deploy a LiteLLM gateway on Google Cloud that routes your Gemini requests with Jev.

• First Steps in AI Solutions Engineering

  Design a C4 diagram first, then build a FastAPI AI service that matches it.

• Triage Incidents with LangGraph

  Build a durable multi-agent CLI that triages incidents with team approvals.

• Build an AI Content Agency with Grok Bot

  Build a five-agent workflow for researched, approved X and LinkedIn campaigns.

Keep the feedback and feature requests coming in our community Discord! https://discord.gg/6ZBmqRK88N

Try NextWork

Happy learning!

Sean Fitzgerald spent 32 years as a paramedic. These days he's a self-taught developer and AI hobbyist in Houston, Texas, and you can count on him to be building something new almost every week.

The list of what he's built so far runs in a lot of directions at once: a social media platform, algorithmic trading bots, a logistics lead scraper, a TikTok automation tool, a RAG system for government contracts, an NFT platform, an AI teaching assistant system, a browser-based chatbot, and his own personal JARVIS, inspired by Iron Man.

The missing 20%

Sean builds almost everything with AI, and he still wants to understand what he's building. An AI answer will take you a long way, but on its own it doesn't give you the technical context, the structure or the practical learning that sits behind it. Sean calls that gap the "missing 20%" - and NextWork projects are how he fills it.

A weekly table in Houston

Sean is now a NextWork Build & Brew master, and his group meets in person every week. You'll often find him at his favorite 24/7 workspace with a coffee, demoing his latest project to the crew.

The crew includes a physician, graduate students, an Uber driver, high school students and career changers - a group with very little in common on paper apart from showing up to build together. They've been meeting for more than six weeks now, often for two to three hours at a time, and they've already completed two hackathons together. If you're in Houston and want a seat at that table, Sean's Build & Brew group is open.

Sean, thank you for bringing a physician, an Uber driver and a room full of students to the same table every week, and for reminding everyone around it that the missing 20% is worth learning properly.

Try NextWork

Nikko is building out of Manila, in the Philippines, and his motto is three words long: Learn, Build, Lead.

When he first joined NextWork, he was an absolute beginner. He was also studying for a Computer Programming degree, and he balanced his projects with that degree the whole way through. From his very first project, he shared everything he built.

Every project, in public

Nikko posted every project he finished on LinkedIn. His early ones were a static website, an analytics dashboard and a chatbot - and on the chatbot, he hit a fallback-response error and debugged it himself before writing it up.

Once he had the foundations down, he kept raising the difficulty. He built a serverless data pipeline that handles more than 100,000 events a day at 98.9% uptime, a computer vision security monitor trained on SageMaker, and a reusable infrastructure-as-code framework that cut deploy time from hours down to minutes.

These days he writes projects for other learners, too, including one on setting up a Hermes AI agent.

From speaker to organizer

Nikko talked about what he was learning from the start, first as a speaker at community events and then as the person running them.

At AWS Community Day Philippines he ran a stall dedicated to NextWork and pitched it more than 1,000 times over two days. As Director of SkillBuilder for AWS Cloud Club Philippines, he built a training program for university students made entirely of NextWork projects. He has been a guest speaker at more than 20 events, demoing NextWork to learners, and before he'd even graduated, he ran a 500+ person event in Manila built around NextWork projects.

Our founder Amber Winton has posted about him on LinkedIn more than once. In one clip, Nikko asks the room, "NextWork Community Philippines?", and the crowd shouts back, "WE LOVE NEXTWORK". In another, she gave him a huge shout out for leading events in the Philippines that use NextWork to upskill hundreds of people.

Still shipping

Today Nikko builds AI and automation systems, runs community events and mentors newer members - and he's still shipping projects of his own.

You can find Nikko on LinkedIn, or catch him at one of his upcoming NextWork events.

When we first shared his story with the community, Nikko said it was "such a great reminder of what we've been through over the years". It's been a great few years to watch, Nikko.

Try NextWork

Zach was 13 when a chef in Greenwich Village handed him a knife and asked him to mince a pile of pearl onions.

He was only in the kitchen because a family friend who imported food from Spain had taken him along on restaurant deliveries in New York City that summer, and had offered the chef a discount in exchange for letting the kid in. The chef watched him work, offered him a prep cook job on the spot, and by the end of the summer Zach was working the line.

From the line to a nonprofit

That apprenticeship carried him through high school and college. The only schools offering him real scholarship money were culinary programs, and he turned them down, because culinary school was meant to get you into the kitchens he was already working in. He went to a liberal arts school in North Carolina instead, and joined the very first cohort of its sustainable food systems degree.

After graduating he went back to New York and into restaurant management, and by early 2019 he was running a steakhouse. Then the pandemic wiped out the industry.

Since he was 11, whenever anyone asked Zach what he wanted to be, his answer had been the same: "I want to own a conglomerate." This is how he remembers that period:

"I had been convinced that I was a failure. Everyone said I was going to go far, that I'd get that conglomerate. Here I was, not even with my own business. And thank God I failed, because I wouldn't be where I am today if I hadn't."

With restaurants shut, he started a nonprofit that gave thousands of pounds of free food to New Yorkers, and ran it full time for a year without ever taking a salary.

A salesman doing five jobs

When they stayed shut, Zach took a tech sales job at a large enterprise reseller that carried more than 200 vendors, as part of the first cohort of its sales academy. He had to learn the whole portfolio at once - and he picked up so many certifications along the way that his resume had to link out to a separate document to list them all.

Startups noticed, and recruited him with the promise that he'd be an Account Executive on day one. What that actually meant at a startup was doing five jobs at the same time: partnerships, channel, engineering, supply chain and recruiting. His title still said sales, but he was already working as a technologist.

His next role was selling cybersecurity products to big banks, at a company that wasn't licensed to sell those products yet. So before he could close a single deal, Zach sourced and onboarded more than 40 cybersecurity vendors to build the portfolio himself.

"This is a 40-hour project"

Zach wanted to understand the technology he was selling properly, so he kept asking the subject matter experts around him to show him how to build a CI/CD pipeline. They kept brushing him off. It was a 40-hour project, they told him, and maybe in a few months they'd have time to explain some of it.

So what did he do? He found NextWork and worked through the full DevOps challenge on his own, and he didn't mention it to anyone. He waited until he was on a customer call with one of those same experts, and at the very end, he dropped the repo:

"And by the way, I made that pipeline. Here's the repo. Mic drop."

That same expert was impressed enough to refer Zach into his next role, and he works in cybersecurity today.

His advice for anyone just joining is about the people around you as much as the projects:

"You came to the right place, and you'll only get the most out of it if you inspire others. This space is free, and it only gets better the more people are here. Bring people in the same way that others brought me in, and you will gain exponentially more value from it."

He means it, too - he now runs his own Build & Brew group, where learners build projects together every week.

Zach, thank you for asking the question until you got the answer yourself, and for bringing so many other learners along with you.

Try NextWork

Shane Brown spent nearly 20 years as an electrician, wiring data centers and managing power systems where one mistake could kill you. He worked his way up from pulling cable to supervising crews, and along the way he got very good at understanding how complex systems fit together.

The travel kept pulling him away from his kids, though, and the stress at home and on the job kept building until he burned out completely. For about seven months, Shane was homeless.

What brought him back was a phone call from his daughter. She told him, in Shane's retelling, "You need to shape up. Be a man. Be a dad." She was working long hours at Starbucks at the time, mostly to pay for his gas so he could come home, and his family rallied to get him back to Arizona. That's where Shane decided to rebuild from the ground up.

Back to the hacker movies

As a kid, Shane had been obsessed with hacker movies and with whatever people were typing on those screens. Once he'd started to heal, he went back to that curiosity. Video game design turned out to be too art-heavy for him, and networking was interesting but didn't grab him, so he used his remaining education benefits to study cybersecurity - in his 40s, and terrified he couldn't compete with sharp 20-somethings. He graduated from Arizona State University with a degree in Cybersecurity.

And then, like a lot of new grads, he couldn't land a job. Every listing wanted hands-on experience that he just didn't have on paper.

Building proof while job hunting

So while he kept applying, Shane started building projects on NextWork, from security labs to web apps, and documented everything he learned as he went. To earn some money in the meantime, he took a job cleaning at a local community college. This is how Shane described it himself:

"A Software Engineer, Web developer and Cybersecurity graduate cleaning toilets and offices. Getting my hands dirty (literally)"

On his second day, Shane was cleaning the marketing department when he overheard that they needed someone to rebuild their website. He spoke up and showed them his portfolio, along with the documentation he'd written for his projects.

On day three, he became the college's new Website Administrator.

What did a marketing team see in someone who had been cleaning their offices the day before? In Shane's words:

"The project documentation is what impressed my new employers."

"NextWork gave me real projects I could show. That's what helped me get hired."

Building for the next person

Shane hasn't stopped building since. He hosted a BuildLab for other learners as part of our 21 projects in 21 days challenge, and he writes and shares resources for people trying to break into cybersecurity and tech from non-traditional backgrounds, which you can follow on his LinkedIn. He talks openly about failure and about rebuilding slowly, and his advice to anyone starting out is this:

"If you want that job, if you want that future, it's waiting for you. But you have to take the steps. One project. One lesson. One day at a time."

Shane, thank you for being so open about the hardest parts of your story, and for hosting a BuildLab so other learners could build right alongside you.

Try NextWork

Dhatri Jayanth is a sophomore at UT Austin, double-majoring in Informatics and Design. She works at the Texas Advanced Computing Center (TACC), she's active in the university's design and product clubs, and she's Adobe Certified in both Photoshop and Illustrator.

Her way into AI tools started at a NextWork event.

The session with Jon

Our founding designer Jon ran a session showing how to use Paper and Claude Code, and Dhatri joined. It was her first taste of wiring AI tools together, and seeing Paper speed up weeks of manual effort was her aha moment with AI tools.

She has kept building since. The thing she's proudest of so far is Synchra, a wearable AR ring she made for a hackathon this year.

Her Build & Brew group

Dhatri is now a Build & Brew master, which means she leads one of our community groups where learners commit to building projects together. Her projects focus on AI-assisted data architecture, and she presents them using AR - which makes a lot of sense coming from someone who has already built an AR ring.

What she wants most from the group is a low-pressure way to get more people building, including people who have never thought of themselves as coders. It's open to everyone with no experience needed, so if that sounds like you, Dhatri's Build & Brew group is a good place to start.

Away from her laptop, Dhatri is a very good oil painter, she has a real talent for organizing chaos into order, and she spends a lot of her spare time reading and cooking. You can follow what she builds next on LinkedIn.

Try NextWork

Derrick Williams started his tech career in the U.S. Army as an IT specialist - imaging machines, running cables and fixing printers. After he left the Army around 2018, he moved through contractor roles and eventually landed a security analyst position in San Antonio, Texas.

It was solid experience. He monitored security alerts, triaged incidents and documented threats, but the schedule was brutal: 12-hour shifts, rotating between days and nights, watching dashboards and waiting for something to break. The quiet stretches were the worst part. When his status update was "everything's fine", nothing was moving, and that included his career.

Starting straight away

Scrolling Instagram on a break one day, Derrick came across a video from Maximus on our team. It asked people to comment if they wanted access to a platform with projects you actually do. Derrick commented, got a response, and opened NextWork to pick his first project straight away.

He started with networking projects, because they overlapped with a university class he was taking and with his security day job. He configured subnets, routing tables, VPCs and CDNs in real environments, with step-by-step videos for whenever he got stuck. IP addressing stopped being a chapter in a textbook and became something he was debugging at 11 PM on a Tuesday.

Networking is still his favorite:

"I would say the networking projects are my favorite. I think networking is an under promoted skill because people like the thrill and title of security/cybersecurity but networking is the foundation of how everything involving cloud operates."

He completed about 20 projects, emailed our team for a personalized roadmap and worked through it methodically. Along the way he passed his AWS certification and started preparing for a security certification, using his projects as living proof of what the services actually did. Somewhere in that process, he stopped seeing himself as "just a security analyst" and started seeing himself as a cloud engineer.

The Muay Thai mindset

When Derrick interviewed for a cloud engineering role, Terraform was one of the requirements, and he was still learning it. So he told the interviewers exactly that, and then laid out his plan to get up to speed using NextWork projects.

He credits Muay Thai for that honesty. In the ring, whatever you're bad at shows up whether you like it or not:

"Either you present your weaknesses or they'll get exposed anyway."

He got the job, and shared the news with us:

"I am excited to share that I have been offered my first Cloud Engineering role. I would like to take this opportunity to thank all the staff at NextWork for not only providing a platform to develop and refine my cloud engineering skills but also for guidance, mentorship and encouragement throughout this journey. Your support has been instrumental in helping me reach this milestone!"

Giving yourself grace

Today Derrick works a 9-to-5 as a cloud engineer, a long way from rotating 12-hour shifts, and he leads a Build & Brew group where learners build projects together each week. His advice for new learners comes from someone who learned around a demanding schedule himself:

"Don't overwhelm yourselves by trying to learn everything at once. And don't buy into the myth that we all have the same 24 hours. We all have high priority responsibilities we have to deal with everyday (Family, Finances, Health etc) so commit as much time as you can afford and give yourself grace."

Derrick, thank you for showing other learners that it's okay to walk into an interview while you're still learning.

Try NextWork

Every week, Niyati sits down with a group of learners for an hour and they build an AI project together. She's one of the leads, and she's learning right alongside everyone else in the room.

That's completely on purpose. Niyati is a firm believer that the best way to learn something new is to teach it - so she co-leads a NextWork Build & Brew group.

The psychology student who loves data

Niyati is a sophomore at UT Austin, studying Psychology and Informatics with a minor in Statistics and Data Science. It sounds like an unusual mix until you hear what she's into: how product thinking and data-driven insights turn into real-world impact.

She's also part of Texas Convergent, where students build their own products, plus Women in Psychology and Women in Data & Statistics. Busy semester!

The projects follow the people

Build & Brew is our community program where learners commit to one hour, every single week, to build a project together. In Niyati's group, the projects follow the people. She shapes each one around what her members are genuinely curious about, and she leads by getting to know everyone in the room.

So what are you in for if you join? MCPs, LLMs, Kubernetes and security projects, and a lot of fun along the way.

Off the screen

Niyati can memorize just about anything, which is a slightly unfair advantage in a room full of new acronyms. Away from her laptop, she's out hiking, at the movies, hanging out with her two dogs, or deep in a good chat with her family and roommates.

If you'd like to learn AI the way Niyati does, come build with her group, and say hi to her on LinkedIn.

Try NextWork

Busola Olugbesan is a full-time cybersecurity student in San Antonio, Texas. She also works part time, leads a NextWork Build & Brew group, and somehow still ships projects every week - probably with the help of a lot of iced matcha.

Building Lunivo

When traditional studying wasn't working for her, Busola started building a different way to do it. Lunivo is the AI-powered learning platform she's building in her own time.

One of her favorite moments building with NextWork was integrating an AI tool into her own workflows, for free. She's been growing in confidence project by project ever since, and she's working toward becoming a Solutions Architect.

Leading her group

Busola is especially good at debugging messy workflows and working out, on the fly, how different pieces of software should fit together.

More than anything, she wants her Build & Brew group to actually feel like a group. That means plenty of talking and plenty of questions, with people helping each other out and building side by side every week.

Good luck with Lunivo, Busola - we can't wait to see it.

Try NextWork

Tristan Cravello came to tech from a middle school classroom. He spent three years teaching social studies - 6th-grade World History, 7th-grade Civics and 8th-grade U.S. History - and coaching track and field.

If you've ever taught, you'll know how that job goes. You plan a lesson, watch it fall apart in second period, and rebuild it before third. Plan, test, adjust, repeat, which as it turns out is pretty much how engineering works too.

More than 80 projects since the start of the year

Tristan joined NextWork earlier this year and has shipped more than 80 projects since. A few of his favorites come straight out of his own life. ClassPulse is a classroom behavior tracker, built by a former teacher. RunLayer recommends running apparel, and Tristan is a competitive distance runner. And threat-ingest is a production-grade pipeline built with Docker and PostgreSQL. You can open all of them in his portfolio.

Running has also taught him that consistency beats intensity, and that setbacks are just part of the course.

Teaching again, every Wednesday

Tristan's whole philosophy is that you can start before you know everything, and he has followed it himself. He co-leads a Build & Brew group, where he writes his own AI projects and teaches them every Wednesday in the NextWork Discord.

He's most drawn to where AI meets security, and you can see it in the directions he's exploring with his group: AI security scanners for CI/CD pipelines, Docker security analyzers, AI scanning for public API endpoints, and Claude Code and Gauntlet Loop-style builds.

For his group he plays a lot of roles at once - mentor, fellow builder, troubleshooter and project guide. He also judges a session the way a teacher would. To Tristan, it's a success when everyone leaves more confident than they arrived.

Try NextWork

Stephen Nimako Jr. was working in technical support in Tennessee, mostly with Azure and Active Directory, when he decided he wanted to move into cloud architecture. He was planning to relocate to Texas for better opportunities, and on paper he was already well prepared, with both the AWS Cloud Practitioner and Solutions Architect Professional certifications behind him.

He was also determined to make his learning visible, so that anyone looking at him could see what he could actually do.

Learning in public

He built projects on NextWork, documented each one and posted what he was learning publicly, so his portfolio grew with every project he finished. His projects went up on LinkedIn alongside his own documentation, which meant anyone curious about Stephen could look at the work for themselves.

He also threw himself into the community. He helped other learners troubleshoot, shared his own Notion project tracker, and ran 30-day challenges where learners completed projects together while they went after jobs.

It wasn't smooth - learning hands-on while switching careers rarely is. Stephen ran into unexpected AWS bills, confusing tools and setup problems, along with long projects, interview nerves and a lot of uncertainty about how to position himself. He turned those experiences into feedback for our team and kept building.

The final stage

Then came a role at Amazon Web Services. By the final stage, the hiring manager was choosing between Stephen and one other finalist. What happened next, in Stephen's words:

"When they brought my résumé to him, he didn't even look at my résumé first - he went straight to my LinkedIn. He picked me over the other guy because of what he saw on my LinkedIn... the projects that I had."

His certifications helped get him into that room. By the time he got there, he also had months of documented projects sitting on his LinkedIn, and that was the first thing the hiring manager went looking for.

Where he is now

Stephen got the role, and the move to Texas he'd been planning happened too. He now works for AWS in Austin.

Stephen, thank you for every 30-day challenge you ran for other learners. You can follow what he's building now on LinkedIn.

Try NextWork

Kemisola is a CRM Manager in Lagos, Nigeria, with seven years in iGaming and marketing behind her. She has built CRM departments from scratch, grown teams and run full marketing campaigns.

The more she used her CRM tools, though, the more questions she had about them. Why does this tool behave this way? Why is this integration so flaky?

"It's a tool issue" wasn't a good enough answer for Kemi, so she went looking for the answer underneath, in the infrastructure the tools run on. She trained her social feeds to show her more of that kind of content, found NextWork on TikTok, and quietly started building, breaking and troubleshooting projects back to back.

Nearly six projects in silence

When Kemisola joined, she didn't introduce herself, she didn't join events and she didn't ask anyone for help. She completed nearly six projects before she posted anything at all.

What finally made her speak up was something she noticed while scrolling through the community: other learners were posting about the same errors, again and again - errors she had already worked through on her own. She recognized the frustration in those posts, and she recognized the hesitation to ask.

"I know what it's like to have questions and not know how to ask. I didn't want anyone else stuck there."

So she started answering questions. Then she started posting her own, and before long she was setting up live calls to walk other learners through the problems she'd already figured out.

Building the Lagos community

Today Kemisola is the NextWork Community Leader for Lagos, a chapter she started with her co-lead Bamise. What began as an ambitious idea now has more than 150 members, and most of it runs through a WhatsApp community where learners build NextWork projects together every week, post them on LinkedIn, hold each other accountable and troubleshoot their errors together in real time.

She does all of this on top of a demanding CRM role, being a mom and her own learning. She often joins live sessions at around 10pm Lagos time, once her kids are asleep and her work is done for the day. Along the way she has shared more than 18 NextWork projects on her own LinkedIn, run more than 10 community events, and coordinated an in-person NextWork meetup in Lagos.

"Don't be a ghost"

For someone who started out so quiet, Kemi is very vocal now about one thing in particular. Her advice to new learners starts with slowing down:

"My biggest tip is to slow down while working on the projects. Don't just aim to complete them - read, learn, and understand what the project is really teaching. Document your projects in your own words, share the lessons you're picking up, and network intentionally. Treat NextWork as more than a task list."

And to every one of those 150+ members in Lagos, she's blunt about visibility:

"Don't be a ghost. If you're not talking about your work, nobody will know you. There are thousands of people with skills… so what makes you different?"

Kemi, thank you for answering the questions you once didn't know how to ask, and for every 10pm session you've joined after a full day of work so that other learners in Lagos don't get stuck alone.

Try NextWork

Roy's first real job was picking up rocks and debris on a property in California.

Today he's a Tech Architecture Manager in San Antonio, Texas, with 16 years in tech behind him. And in less than a year on NextWork, he has completed more than 200 projects - every single one of them is sitting in his portfolio for anyone to open.

By almost any traditional measure, Roy has already made it. So why is he still learning at this pace? Because the job he worked so hard to get keeps changing, and building is how he has learned every single thing in his career so far.

Three careers and about twenty jobs

Roy didn't take the straight road into tech. He worked his way up in construction until, in his words, his body made that plan stop working. Then he moved into hospitality, and he chose it on purpose: he couldn't talk to people, so he took a job that would force him to. That became physical security for hotels and condos, and then an operations manager role.

His way into tech was a support desk job, with no tech experience at all. This is how he describes what came after:

"I bridged the gap with what hospitality had taught me, and it was still a completely different game. Three careers and about twenty jobs later, I am a senior engineer and a manager. All of it self-taught, on the job, by building.

Would I change any of it? No."

"Another way of learning"

When AI came along, Roy did the same thing he'd done with every career before it.

"I thought AI was for people with PhDs, locked in a lab somewhere. It is not. It is another way of learning, and you get at it the same way you get at anything else, by building with it."

He calls himself a "lazy engineer", and he builds accordingly. Most of his projects are systems that do the manual work for him, so he has more time for his family, his health, real estate, his hobbies and the advisory business he runs on the side. One of them is a mock interview system with three AI personas - a hiring manager, a technical lead and HR. Another is an Excalidraw-to-Linear workflow powered by Obsidian, Codex and 25 agents.

He says building these has changed how he works, both in his day job and in his own businesses.

Teaching what he learns

Roy believes you lock in what you've learned by building real things and then teaching them to someone else. He sums it up as "prompt, build, connect", and it's why he now runs a weekly Build & Brew group, where a small group of learners meets to build projects together.

He is very clear that he isn't there as the expert:

"Am I an expert? No. I am 10+ months in and building every day. That is the whole point of this Build & Brew group: I am a little further down the same road you are on, and we cover the rest of it together."

Our founder Amber Winton first wrote about Roy in The market most VCs miss, and he's still one of the clearest examples we know of someone who never really finishes learning, because the work he's responsible for never really stops moving.

Roy, thank you for showing up for your Build & Brew group every single week, and for proving to so many people that you don't need a PhD to build with AI.

Try NextWork

🐛 Bug fixes

• Answers in a project’s validation boxes now save a second after you stop typing, so a page reload halfway through won’t wipe them.

• People search wasn’t showing anyone who joined after 19 September - we’ve added a fix so all members show up now.

• In dark mode, every checklist box inside project steps used to appear ticked, so steps you hadn’t started looked finished. The boxes now stay empty until you tick them.

• Searching a project’s full title used to bury it under similar projects, so you had to scroll to find it. Now that project comes up first.

  

🚋 In progress

• Reopen your past chats when you generate a project, so you can pick up where you left off.

• Smoother sliding when you switch between topics in Explore.

🧪 New projects

• Build a Custom Agent Harness

  Build a TypeScript agent harness with tools, approvals and OpenRouter, and see what really happens between your prompt and the tools a coding agent runs.

• Build a Resumable LangGraph Workflow

  Build an incident workflow with LangGraph that pauses risky actions for human approval, then picks up again from where it stopped.

Keep the feedback and feature requests coming in our community Discord! https://discord.gg/6ZBmqRK88N

Try NextWork

Happy learning!

This week, your portfolio got a Work page and footer links, you can remix any project into your own, learnlists became searchable, and dark mode arrived across NextWork.

💼 Your Work page is live

Your portfolio has a Work page that shows every project you've finished or are working on, so anyone you share your portfolio with can see what you've built.

🔀 Remix any project

Open the ⋯ menu on any project and choose Remix project.

It opens that project in a new draft and asks you a few questions to make it yours, so you start from something that already works instead of a blank page.

🔎 Learnlists are searchable

Public learnlists show up when you search, with their own Learnlists tab in the results. Search for a topic and you'll find the lists other learners have put together for it.

🔗 Add your links to your portfolio

You can now add your GitHub, LinkedIn, X or your own site to your portfolio footer, so anyone who likes your work knows where to find you.

🌙 Dark mode is here

Dark mode has been one of the most requested features in our community Discord, with a thread all of its own. Community member Zach Gruenberg even built his own NextWork dark mode extension for Google Chrome back in August. Thank you, Zach, and thank you to everyone who asked for it!

Explore, project pages, roadmaps and portfolios (… everything!) all come in light and dark, with illustrations made for each.

✨ Smaller improvements

• Students have their own sign-in tab. Choose "I'm a student" and sign in with your school email.

• Links to a missing page, or to a private project or learnlist, open a page that tells you why it's missing or private, with a way back to projects.

🐛 Bug fixes

• Opening a NextWork learnlist used to hide the Explore bar. It now stays in place.

• Long messages in the edit chat sometimes came back with an error. They go through now.

• Some project pages crashed on load in certain browsers. They open normally now.

🚋 In progress

• More ways to browse your portfolio, including list and grid views.

• A new setup flow for the desktop app that asks for your name and writes a first draft of your About blurb.

Keep the feedback and feature requests coming in our community Discord: https://discord.gg/6ZBmqRK88N

Try NextWork

Happy learning!

Look at how the UN is set up: the World Health Organization for health, the World Food Programme for hunger, UNHCR for refugees and the International Labour Organization for work. None of them is organized by profession. Getting food to people in a crisis takes logistics, nutrition, engineering and economics, so the World Food Programme brings all of those people together around the one problem.

Education works the other way around. We sort learning by subject: majors, departments, degrees, and eventually job titles. Most of us pick a subject first and work out what it's for later.

You can hear it in the questions students ask: "What major should I do?" rather than "What problem do I want to solve?", or "Will this be on the test?" rather than "Will this make a difference?" A lot of that comes from the advice we give them. If a 16 year old is determined to get plastic out of the ocean, we tell them to focus on their grades and wait for a marine biology degree, when they could be prototyping a solution today.

How learning got separated from problems

This wasn't always the case. For most of human history, learning and survival were the same thing. If someone in the tribe ate a berry and got sick, everyone learned to stay away from it and made sure their families knew too. What you learned kept you alive, and I think of that as learning for survival.

Then the industrial age put a job between what we learned and why it mattered, and with it came a manager, a ladder and a title to collect. Getting ahead started to depend more on where you sat in a company than on the value you added, so we learned to collect credentials instead. I think of that era as learning for status, and it's still where most of us are today. It worked while a degree reliably led to a job and then a career, but technology now moves faster than any curriculum can be rewritten, and the problems don't wait for the credential to arrive. Climate change isn't pausing while the next generation finishes an undergrad degree.

Climate change is one problem that no single subject can solve, and an aging population is another. Solving it needs architects to redesign cities, engineers to build assistive technology, psychologists who understand isolation and economists who can rethink retirement. Today those people sit in different faculties and companies, and they rarely meet until they're well into their careers - if they meet at all.

The UN figured out long ago that a problem like that needs one team built around it, and education still hasn't caught up.

The UN builds its agencies around problems like hunger, health and work, but when leaders talked about skills last week, they mostly talked about them on their own. Singapore called for investment in infrastructure and skills, Jamaica for infrastructure, skills and capacity, and Zambia for education, skills, technology, entrepreneurship and decent jobs. They're right that skills matter, but a skill is most useful once you know which problem it's for, and that's how the UN already organizes everything else.

What changes if we start with the problem

What would it look like to organize learning around problems?

Imagine a Department of Access to Healthcare instead of a School of Medicine, where doctors, nurses, engineers and economists all work on getting care to people who can't reach it. A student in that world wouldn't say "I want to be a doctor", they'd say "I want to save lives", and then learn whatever that takes.

Starting with the problem doesn't make skills any less important. It changes the order you learn them in, and that's what I'd call learning for solutions. It's the reason we built NextWork. Our mission is to meet the skill demand to solve the world's toughest problems, and every project on NextWork starts with something real that anyone with a laptop can build.

That 16 year old determined to get plastic out of the ocean doesn’t have to wait for the degree to start working on it. They could begin this afternoon.

Try NextWork

All my best,

On September 22, Andreessen Horowitz announced The Horowitz Andreessen Academy, a new school in San Francisco for people straight out of high school. What was most interesting to me was that students are chosen on proof of work: what they built, shipped, or earned. Transcripts and test scores are still collected, but the FAQ says admission is "heavily weighted toward proof of work."

To anyone who has ever hired an engineer, choosing on proof of work sounds obvious. Of course we look at the work, and most of us do it the moment a candidate sends through a GitHub link. A project someone built tells us more than lines on a resume ever can, because it shows us:

• how they thought through the problem

• what they chose to build, and what they left out

• what they did when something broke

That matters even more now. AI makes a polished answer easy to produce - a cover letter, an essay, a neat written response to almost any question - but a working project, with decisions the builder can explain, is much harder to fake. The Academy's own FAQ says it plainly: "What we want to see is how you think and what you've actually built."

So is the Academy right to ask for the work? Yes, and I hope more schools and companies do this.

We built NextWork around the idea that proof of work beats credentials. Now one of the most influential firms in tech is choosing students on the same basis. That's good news, and it raises a harder question.

Choosing people on proof of work only helps the people who already have some proof to show. So who has proof of work to show? What did those applicants do before they applied? And where did they get the chance to make it?

The Academy is honest about the answer. It describes itself as being "for a specific kind of student: someone who already knows how to build," and its launch release goes further, admitting that "college is still the right path for most students." That leaves the bigger job to the rest of education, which is helping everyone who doesn't have proof yet go and make some.

Dropping the degree requirement was never enough

The idea that a credential shouldn't decide who gets in is not new. In 1971 the US Supreme Court decided Griggs v. Duke Power. The company required a high school diploma to move out of its labor department, the lowest-paid part of the business. Chief Justice Burger's opinion put it this way:

"Diplomas and tests are useful servants, but Congress has mandated the commonsense proposition that they are not to become masters of reality."

Put plainly, a credential shouldn't decide who gets in when it doesn't measure whether someone can do the work.

Fifty years later, plenty of employers took that idea seriously and dropped degree requirements from their job ads. Then researchers checked whether it changed who actually got hired. In February 2024, the Burning Glass Institute and Harvard Business School found that it barely did: across the US, dropping the requirement changed the outcome in "not even 1 in 700 hires last year."

To me, that says removing a requirement on its own doesn't do much. Whether the ask is a degree, a certification or proof of work, people still need real experience, and evidence of that experience, before anyone will let them in.

And the place most of us got that experience, the entry-level job, is getting harder to reach. Stanford's Digital Economy Lab tracks 22 to 25 year olds in the jobs most exposed to AI, and in August 2026 it found their employment had dropped by about a fifth compared with people the same age in less exposed jobs. Experienced workers in the same jobs saw no such drop.

Fewer entry-level jobs means fewer people shipping their first real thing, breaking it and fixing it, with a manager who can vouch for the work. So fewer people have anything to show when a school like the Academy asks for it.

What NextWork does

NextWork doesn't select people who already have proof of work; we give people a place to make it.

Every NextWork project is real work, done with the tools professionals use, so when you finish one you have something you built and can explain, and a record of what you're capable of. The point is to have that record before an employer, a school or any other gatekeeper asks you to show it. More than 220,000 people in 190 countries are building theirs on NextWork right now.

If proof of work is the new credential, then the job of education changes with it. It has less to do with sorting people by what they have already built, and more to do with giving someone with nothing to show yet a place to make something.

Proof-based selection is the right idea, but it is only fair if far more people get a real chance to make proof before someone judges them on it. Right now that chance mostly goes to the people who already know how to build - and widening it is the work that matters.

Try NextWork

All my best,

There's a new way to discover projects on NextWork.

Our Explore menu now introduces you to individual projects first, so you're seeing the latest and trending projects right away.

You can also browse projects by topic. Pick something you want to learn, like DevOps or Security, and you'll see the roadmaps for that topic first, followed by Learn Lists for going deeper.

New to the search bar is that you can now also search people! Look up a name and you'll find learners, alongside the awesome projects they've created.

Last, but definitely not least, we've added new starter ideas for creating your own project, including AI DevOps, MCP and building an agent harness.

✨ Smaller improvements

• "Your Library" in the toolbar is now just "Library".

• The Update button on the edit page is now called Publish, to match the Publish window it opens.

🐛 Bug fixes

• The progress circle on a project now looks the same in Explore and in your Library.

• On phones, the toolbar at the bottom of the screen used to sit on top of the home bar, hide behind the keyboard and cover the bottom of every page. It now stays clear of all three, and its search box is full width.

• Some Explore shelves showed 0 completions for older projects, even when learners had finished them. Shelves now show the same count as the project page.

• Profile pictures had a dark brown ring around them. They're back to a faint grey outline.

• On phones, project rows cut the category off mid-word and pushed the "completed" count off the screen. That line now wraps instead.

🚋 In Progress

• A Work page on your portfolio that shows every project you've finished or are working on, with a live cover for each, rolling out this week

• A faster Explore homepage that loads instantly instead of waiting on the full catalogue

• More ways to browse your portfolio, including list and grid views

Please continue to share your feedback and feature requests in our community: https://discord.gg/6ZBmqRK88N

Try NextWork

Happy learning!

Before people are trusted with bigger decisions, they usually practise with the smaller ones: cleaning up a spreadsheet, fixing a bug, writing a first draft, responding to a customer’s support request… you get the picture.

From the outside, it looks like grunt work. Much of it is slow, repetitive and boring. But some of those tasks give beginners a chance to make a call, see what happens and revise their work while the stakes are still manageable. With useful feedback, they learn what to do differently next time.

When we delegate that work to AI, we lose those opportunities along with the tasks. The result arrives ready for approval, but someone who has never worked through a similar problem doesn’t know what good looks like. Reviewing requires experience they don’t have.

This is not a plea to keep junior drudgery, but asking people to approve work they have not learned to evaluate creates a problem education needs to take seriously. We cannot tell people that their value now lies in judgement while taking for granted how we’ve been developing it all this time.

What this means for NextWork

If fewer opportunities to develop judgement come through junior work, learning experiences need to provide more of them. But simply giving someone a project does not solve the problem. A project can make every decision for the learner just as easily as an AI tool can.

For NextWork, this means being deliberate about which parts AI handles and which decisions the learner needs to practise. Those choices should depend on what the person is learning. Something worth working through on a first project may be sensible to automate on the next, leaving them free to tackle a more demanding problem.

We do not need to preserve every task AI can now do, but we do need to preserve the opportunity to learn from decisions that matter, then move people towards work that demands greater judgement. That is what NextWork provides with hands-on projects, at every stage of a learner’s career.

Try NextWork

All my best,

Ask an AI how to become a cloud engineer or an AI engineer and it can produce a credible roadmap almost instantly: courses, tools, projects, certificates, timelines and milestones.

Ask again and it will often produce a different version.

Neither answer is necessarily wrong. There are several reasonable ways into a new field, and the value of AI is that it can make those possibilities available to more people. The problem is that a learner is now left to decide which route is worth their time, even when they do not yet know enough about the field to make that judgement confidently.

Education used to make the choice for you

Traditional education made that decision on their behalf. You enrolled in a course, followed the sequence, and trusted that somebody had thought through what should come first.

There were obvious limitations - the path could be outdated, too generic, or completely wrong for why you were learning in the first place. But it did remove the burden of deciding what to do next.

AI has created the opposite condition. Learners now have access to more explanations and possible routes than anybody could follow, so the difficult part is no longer finding something to learn. It is working out which next step will actually move them forward.

That decision cannot begin with a subject alone.

Someone may say they want to learn cloud, AI or Python, but they may be looking for better work hours, a way into a new industry, people who understand the work they want to do, or something concrete they can show when they apply for the next job. Someone else may be trying to automate a task that is making their current job harder than it needs to be.

A roadmap needs a destination and a location

At NextWork, we think motivation is the reason somebody starts, while inspiration is the future they can picture themselves reaching. Together, they give a learning path a destination. To recommend the right next step towards their destination, we also need evidence of where the learner is now.

That is what projects give us. When somebody builds something, we can see what they wanted to do, what they were able to complete, where they got stuck and what they may be ready for next. Instead of asking a learner to choose between beginner, intermediate and advanced, or asking AI to produce an entire route before it knows them, each project becomes evidence for the next recommendation.

A good roadmap should not be a fixed plan handed to somebody at the beginning of their learning. It should become more useful as they build: anchored in the future they want to reach, adjusted by the work they have already done, and clear about why the next project is worth their time.

This is why we have built NextWork around projects, rather than a fixed catalogue of content. Projects move learning from an abstract interest to real work. They give learners something they can share with a community, add to a portfolio, or show to an employer, while giving us a much better basis for understanding what should come next.

What NextWork is building towards

The work ahead is connecting those moments without putting learners back into a fixed curriculum. We want a learning experience that is always learning why the learner is here, what they have already done and where they are trying to go, then uses that context to help pick the next project or change direction if the facts change.

The difficult part is not creating more roadmaps, AI can already create more routes than anybody can follow.

It is helping somebody choose the next piece of work with enough confidence to begin. That’s the real opportunity around roadmaps we’re excited about: building the layer that turns a learner’s motivation and the work they have already done into a clearer answer to what they should do next.

Try NextWork

Happy learning,