Adam Bossy-Mendoza
Brooklyn, New York, United States
2K followers
500+ connections
About
See my writings to get a sense for how I think at adambossy.com — I'd love your…
Activity
2K followers
-
Adam Bossy-Mendoza shared thisMy experience interviewing earlier this year was counter to what everybody believes is the reality of an extremely harsh tech market. I went from first recruiter call to signed offer in 23 days and had 15 companies in my pipeline at the peak. It felt like a bonanza. I’ve heard what’s going on right now described as a “massive redistribution of talent” due to AI. That’s an understatement. We went from ZIRP-era hiring bloat to what feels like amazing teams being assembled around extraordinarily ambitious ideas. There are no 1-person unicorns yet. That’s just a VC fantasy. Startups with traction are hyperscaling as always, though perhaps at a slightly slower rate. The felt demand for talent was massive to me and felt like 2021 levels. 🧠 The Technical Interview Irony Interviewing, unintuitively, feels harder than ever. When I started my career, there was a collection of leetcode-style questions that were industry standard. Sure, they weren’t related to our day-to-day work, but they were a known quantity. We could practice them as an independent skill and the questions in the interview would match our expectations. They were hard, but actually sorta fun. In the 2010s, there was a cultural shift in the industry to align interview questions with the job’s actual work. Intuitively, this makes sense, but I hate this shift. It’s resulted in companies asking questions about precisely what they’re building. Given the interviewer lives and breathes this problem every day, and everybody they work with does also, this imposes a *huge* information asymmetry. When we were all asking leetcode-style questions, we were mostly on a level playing field. The interviewer likely knew their particular question well, but was equally as bad as you at leetcode. To boot, companies are now layering everything together. To prepare today, we have to: 1) Know leetcode in and out. 2) Know the company’s own architecture as well as they do. 3) Be prepared for unpredictable AI questions. 4) Study system design The breadth is extraordinarily wide, making effective preparation feel elusive. 🧘♂️ Managing Interview Anxiety I started my search feeling really anxious and freezing during coding screens. During a well-timed session, my therapist advised me to “surrender and trust the process.” It was exactly what I needed to hear. I start passing all of my coding screens. I actually *stopped* practicing at that point and started doing better than before. 🎲 Embrace the Randomness Interview decisions are completely arbitrary for reasons that you can’t control. This is true for myself and for star performers I’ve worked with. People aren’t expert judges at predicting whether a person they’ve met for 60 minutes will succeed in a complex, evolving role for years to come. It’s a safe bet to assume that, maybe 50-80% of companies will tell you no and that you simply need enough volume to get 2-3 high-quality yeses. What has your interview experience been like in the AI era?
-
Adam Bossy-Mendoza reposted thisAdam Bossy-Mendoza reposted thisEveryone building a company eventually says it: "Things will get easier once we reach X." They won't. And that's actually the point. Building a company is like playing a video game. At each level, you earn the right to face a harder boss. The difficulty doesn't decrease - it scales with you, and you earn that with each win. When you close your seed round, you unlock harder fundraising. When your product works, you unlock stiffer competition. When you build a great team, you unlock more complex leadership needs. Every win is permission to attempt something more difficult. The trap is thinking success will eventually deliver relief. It delivers opportunity, which usually looks like more complexity, higher stakes, and new unknowns. But here's the thing about video games: if the difficulty never scaled, you'd never get to use what you learned by playing. The whole point of leveling up is that the next level tests your new capabilities. You practiced the sword for a reason. The same is true whether you're taking on technical debt, leading your first team, managing a new product line, or facing real competition for the first time - the market taking notice of you is a boss level, not a finish line. Stop waiting for easy. You're playing the game. And you're getting better at it.
-
Adam Bossy-Mendoza shared thisI may have accidentally created the world's most powerful personal finance tool 😅 What started as a project to vibe code a better, AI-powered Mint.com for myself eventually became both a simpler and much more powerful tool to analyze my personal finances and help me budget. Each time I wanted more capability and felt lazy, I handed more control to AI and pushed into more of an agentic model. What came out of it was a personal finance agent I'm calling Transactoid. My wife and I did a deep dive on our finances using Transactoid and the power of the insights it yielded were palpable. We just kept going deeper and deeper, asking question after question. This simply wasn't possible before. It runs locally, and if you need help setting it up, don't hesitate to reach out. Feel free to improve upon it also—I'd love it if somebody enabled it to run in the background and yell at me when I go over budget. A detailed write-up and link to the repo are below 👇
-
Adam Bossy-Mendoza reposted thisWe love to see it. 📈 Watch how GPT-5.2 gives Charlie the control and context it needs to ship complex PR chains: 5 issues, 5 PRs, 1 clean merge path. Thanks to the Charlie Labs team for sharing!Adam Bossy-Mendoza reposted thisWe’ve been testing GPT-5.2 for the past week, and it just set a new high-water mark on our internal evals. It’s exceptionally strong at following precise instructions across complex, multi-turn agentic workflows—especially with large amounts of context. Here’s Charlie running on GPT-5.2 👇🏻
-
Adam Bossy-Mendoza shared thisSee Charlie work exceptionally well with GPT-5.2, released today!Adam Bossy-Mendoza shared thisWe’ve been testing GPT-5.2 for the past week, and it just set a new high-water mark on our internal evals. It’s exceptionally strong at following precise instructions across complex, multi-turn agentic workflows—especially with large amounts of context. Here’s Charlie running on GPT-5.2 👇🏻
-
Adam Bossy-Mendoza shared thisCharlie Labs is building the world's most advanced coding agent. We're hiring for 3 roles to join our very talented, small and distributed team: * Product Engineer * AI Engineer * DevRel If you'd like to take a big swing at building what the future of coding can become, please reach out!
-
Adam Bossy-Mendoza posted thisAfter 2+ years of AI engineering, my prompts consistently end up in utter chaos. They end up randomly strewn all across my codebase, changes get buried in git history and it becomes hard to track which prompt produced which output in a systematic way. I often end up wasting valuable cycles doing the bookkeeping just to keep everything organized. Promptorium solves this. It versions, stores, and organizes your prompts as you iterate, independent of Git, so prompt changes don’t disappear into commit noise. It helps your prompts evolve in a transparent, intentional way, making them easy to reference and improve for evals, A/B tests and other analytics. WHAT IT OFFERS - Reproducibility: Know *exactly* which prompt version generated which output. - Explicit, reliable history: Refine prompts without losing the evolution. - Organization by default: No more scattered `.md`, `.yaml`, `.txt`, and inline strings. - Separation of concerns: Prompts live apart from deterministic code. - Developer ergonomics: Remove chaos; gain clarity, structure, and consistency. HOW IT WAS BUILT A very cool thing about this project is how the library came together. I had the architecture building in my mind for months, and after I finally sat down with an AI coding assistant to firm it up, the components came together quickly. I gave those pieces to GPT-5 Pro and about 45 minutes later, voila! It had one-shot the implementation, with the code quality you see today. It's well-decomposed and easy to reason about, unlike a lot of AI-generated slop, and roughly scoped to my liking. Before AI, I would have let my opinion of the architecture evolve as I wrote the library, but now, it's essential to lock a great architecture in place first, investing the upfront time to shape it. It shows how providing AI with the right boundaries can lead to a superior result, since without AI, I probably would never have had the time to write it, let alone make it neat enough to publish. I focused on the architecture and design level and trusted that the AI would handle the implementation. Ultimately, I did the thinking, making a lot of tiny decisions along the way that AI would have gotten wrong and trusted it to do the rest. WHAT'S NEXT - A full TypeScript port: parity across Python and TS codebases. - AI Refinement: modify your prompts based on real data, using itself for its own prompt bookkeeping. What could go wrong? Link below 👇
-
Adam Bossy-Mendoza posted thisI’ve wanted to build this tiny utility for over 15 years, but it never felt worth the effort until now. Like most developers, I’ve got a handful of dotfiles (my .bashrc, .vimrc, and friends) that evolve naturally and somewhat chaotically as I try new tools and iterate on my setup. Over time, they grow annoying to track and they fall out of sync across machines. Inevitably, they break or I have to start over when I get a new machine. I finally built dot-sync, a tiny bash utility that keeps my development environments tidy and consistent. Plot twist: it was written perfectly by AI. What dot-sync does: - Syncs dotfiles between your local system and GitHub - Pulls updates back down cleanly - Shows diffs before applying any changes - Confirms every action before syncing - Makes it almost impossible to lose track of your setup again The trickle-down effects of AI This type of mundane problem is far from the bullseye of what AI labs are trying to solve. AI made it trivial to write, debug, and enhance something I’d postponed for over a decade. I just couldn’t have justified the time to dig into bash to achieve this level of quality. Not only did it get the shell logic right on the first pass, it even added features I wouldn’t have bothered to code myself. It dramatically improved my developer experience as a result, and it’s led me to iterate on this about two-dozen times since initiating it. Why it matters It’s a small project, but it’s made a big difference in how I work. If you’ve ever lost your favorite shell alias to a new laptop, this might save you some pain. Link below 👇
-
Adam Bossy-Mendoza liked thisAdam Bossy-Mendoza liked thisBuyers... be liars. Turns out they don't lie to agents. Last night, while everyone here slept, a buyer spent 9 minutes with our AI demo agent. Asked every question they had. The stuff buyers usually hold back until call three. Then they told it exactly what it would take to win their business... and booked time with a rep for tomorrow. The call hasn't happened yet, but we already know their pains and what winning looks like. The roadmap to close... Here's the part I can't stop thinking about. They were more open with the agent than buyers are with a 'human'. No fear of being pitched... no performance.... nobody on the other end trying to close them. So they just... said the truth. Gasp. No pitch anxiety.... no just looking.... no holding the real answer hostage until you earn it. The guarded buyer isn't guarded. They just don't trust us. Chew on that one. While you look at a drone view of the sunrise, over the Hudson...
-
Adam Bossy-Mendoza liked thisAdam Bossy-Mendoza liked thisExtend's Q2, by the numbers: → 6x YoY growth (8-figure ARR) → 1B+ pages processed → 280% NRR → team 3x'd since January → 2 major launches shipped every month → We closed the largest deal in company history (head-to-head against all of our competitors) Some customer quotes from the quarter: "I've worked with a lot of dev tools in my career. This is the best dev tool I think I've ever worked with." "Of all the vendors we work with, Extend is the one we worry about the least. You guys are definitely critical infrastructure for us now." "I completed the evals for Gemini and the evals for Extend... this is the difference between hell and heaven." "I signed up for Extend from a personal side... I was actually blown away. Absolutely blown away." "Our ops team is like... my mind is blown. How is it so good at doing this?" "I only looked at you guys and [redacted]... the demo from you guys was miles better." onwards to Q3!
-
Adam Bossy-Mendoza liked thisAdam Bossy-Mendoza liked thisWhat My Father Taught Me My father passed away last week. He was 102! He and my mom were married for 78 years (she's 101). People keep asking what his secret was. He'd say bourbon. Hard to argue... But watching him, I think it was something else — he just never let life stress him out. Not in a checked-out way. He felt things, cared about things. He just took them in stride, always, and somehow kept his footing no matter what came. In his late 80s, back surgery paralyzed his dominant arm. He taught himself to write and eat left-handed and never once complained. He grew up a few blocks from the Golden Gate Bridge and walked across it on opening day in 1937. That same curiosity took him a lot further, eventually — 110 countries, often decades before it was common to go. And he showed up for people quietly, without needing credit: a family that arrived in California as refugees became lifelong friends, given a seat at our Thanksgiving table for years. Their son wrote to us this week: "You showed us that real success is measured by the love of a devoted family... and the belief that there is always room at the table for one more." And every night when I was small, he and I had one scoop of ice cream on a TV tray, in the room we weren't supposed to eat, watching whatever we were into that week. I still have ice cream five or six nights a week. Now you know why. At my 50th birthday, he told me: "It's not the years of life that count, but the life in your years." He somehow managed both. I don't have a tidy lesson to close with. Just gratitude — for 102 years, fully lived, and for having been his son.
-
Adam Bossy-Mendoza reacted on thisAdam Bossy-Mendoza reacted on thisThis Fall, Thesis Driven is taking a new form: a 60-page, glossy print magazine. The reason might surprise you. It’ll include feature articles, operator profiles, original reporting, and editorials from me, Paul, and Thesis Driven's top guest contributors. Mailed anywhere in the US and Canada. Here's why we're doing it. As we studied how people read Thesis Driven, we discovered something special. People didn't consume it when it arrived in their inbox. They waited for letters to pile up, then binged on flights or Saturday morning, much like they once read the paper. Print fits that behavior better than a feed ever could. There's a second reason, and it's even bigger. Digital channels are saturated. AI sales tools like Clay, paired with richer and richer contact databases, let any inside sales rep fire off thousands of "custom" emails a day to anyone with a title. LinkedIn, X, Instagram are heading the same way: bots talking to bots. As AI content gets harder to tell apart from human work, there's a real chance the internet starts losing media market share. At least among the specific, sophisticated audience we're trying to reach. Our digital reach is bigger and faster-growing than ever. But this isn't the time for media brands to rest on their digital laurels. We think the magazine is a fun, thought-provoking read on its own. But we also can't think of anything better to leave on the coffee table in your lobby that says you know what's what in real estate. This is the first of many. We're publishing biannually. Link to grab a copy is in the comments.
-
Adam Bossy-Mendoza liked thisPaul Gleger and I are building something exciting inside DriveX Labs. All the coolest leaders are nerding out about the tools and products that they used to only dream about and now can build on their own. But "on your own" is lonely, no one is there to show off the cool thing that you built, no one but Claude to ask your questions to. We want to build the community behind the builders. A curated space of senior leaders who are getting back to hands-on work. A place to show off, share ideas, and build together.Adam Bossy-Mendoza liked thisEmerging challenge for engineers: loneliness
-
Adam Bossy-Mendoza liked thisAdam Bossy-Mendoza liked thisOne of the weird side effects of there still being a huge WFH/remote work culture is that it seems even harder to show the younger generation what different jobs are really like and inspire them to achieve more. When I was growing up, I would visit my parents' (and sometimes, friends' parents') offices and see such a variety: - A bookkeeping firm office that was very plain with tons of paper shredders, older computers, and binders everywhere. - A video game company office that had ping pong tables, a conference room with all the latest game consoles, posters of the games they've published on the wall, corner offices with views - A tech company office with empty hallways but large offices per person. - An administrative office with nothing but cubicles. - A small business office with a giant stuffed bulldog in the lobby that people would put hats and pins on, scooters, and themed conference room - An architect office with scale models along the wall and boards with blueprints and drawings of what the building will look like As a kid, I didn't really understand what any of those people did, but I could at least tell you which vibes I liked best and which people in the office I thought I might like to be like. Nowadays, ALL of those jobs can be done remotely, and largely are. My niece and nephew's parents and their friends' parents mostly just work from home, so all they think white collar work is is going on Zoom and having meetings all day at home. I'm happy that people have the flexibility to work from home, especially when they have kids to raise, but I also wonder about the types of side effects remote work has had on kids that we don't talk about much. The "take your kid to work" day hits different today.
-
Adam Bossy-Mendoza reacted on thisAdam Bossy-Mendoza reacted on thisAn update! I am joining Anthropic as a Member of Technical Staff! I’ll be on Labs -- it’s a founder shaped role and I couldn’t be more excited to have the opportunity to help shepherd technology for the benefit of humanity. Onwards towards helpful, harmless, powerful AI.
-
Adam Bossy-Mendoza reacted on thisAdam Bossy-Mendoza reacted on thisIntroducing Layerize. You'll never design images the same after this: Our intern Aditya Mehta just shipped this and it blew my mind the first time I tried it. You can now drag in any image and Moda rebuilds it into editable layers - text, elements, all of it - ready to move, retype, and remix directly. And the text reproduction holds up better than Canva or Photoshop, according to early users. Here's some ways to use it: 1. Generate with ChatGPT or Gemini → Layerize in Moda → Directly tweak small details 2. Find inspiration image → Layerize in Moda → Remix to match your brand 3. Grab old photo → Layerize in Moda → Clean up extra noise Comment "Moda" for access.
Experience
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Education
-
The University of Texas at Austin
-
-
Activities and Societies: Association for Computing Machinery, Hispanic Business Student Association
Winner of Outstanding Recruitment Program among international Association for Computing Machinery chapters while I was President. Led organization of 50 active members.
Volunteer Experience
-
Mentor
Hackbright Academy
- 4 years 4 months
-
Technical Project Lead
YCore
- 7 months
Economic Empowerment
Honors & Awards
-
Toyota Scholars Program, 2005-2008
Hispanic Scholarship Fund
Languages
-
Spanish
-
Recommendations received
1 person has recommended Adam
Join now to viewOther similar profiles
Explore more posts
-
Elizabeth Zalman
I am super excited this morning to announce Sandgarden's acquisition of Layer! When I first met Jonah Katz and Andrew Hamilton, I fell in love with their enthusiasm and positivity, not to mention what they were building. Two years later, I'm proud to have them join Sandgarden. Thanks to Michael Hirshland for the introduction, and to Andrew and Jonah for believing in what we're building to take a chance. Congrats on getting it over the finish line! Welcome to the team Ayaan G. and Gavyn P.! Finally - check out Doc Holiday! It keeps everyone in the company up-to-date as features ship, starting with release notes and documentation all the way to team-specific Action Plans. Demo video here: https://lnkd.in/grkAuXkh And finally - a shout to Jonathan J. Chou, Nicholle Lamartina, and Tori Stark for the insane help. You never fail to deliver 🥰
12 Comments -
Karam Aboul Hosn
BeReal sold for $500,000,000. Me and Omar Zeineddine think we can build something bigger. We aren't the most cracked devs nor do we have funding, but we do know what BeReal got wrong and that is our leverage. Today we sat down with 6 of our beta testers and just listened. What stood out to us the most wasn't even about features, it was about behavior. Here are some things that they told us they like about using Vlogit, Inc.: -They like how you don't just see a moment in your friends day as you would with just a photo, but rather you get to see their whole day - How its only targeted to their friend groups and they dont have to worry about other people watching or if their posts are "post worthy" - They love how you can always come back and rewatch you and your friends vlogs - They also said the app has motivatived them to do more with their days so they can show their friends Now more importantly, here are some of the things that they didnt like as much: - The biggest thing was that there isnt much to do on the app right now when you are not picked to vlog - When mentioning a public feed, they didn’t find see themselves watching that many vlogs from strangers unless it was a curated feed (i.e., a feed built for their own preferences.) Those parts matter more than anything. BeReal proved that people want something more authentic, but they failed to adapt to shifts in user behavior. We plan on listening to our users to make sure that doesn’t happen.
-
Steve Vassallo
I’ve come to believe that early experiences carry disproportionate weight - far more than we often acknowledge. For designers and builders, “minimum viable” may suffice for validating core functionality, but it rarely earns trust or sparks genuine delight. While velocity matters, I’ve always struggled with the notion that early versions should be merely viable, especially in a world where attention is scarce and emotional resonance is everything. This extends beyond product. Do you show up as the minimum viable version of yourself in your work, your relationships, your craft? I would encourage you to challenge the conventional MVP framework. Replace it with the Minimum Awesome Product. This doesn’t mean it’s fully-featured or pixel perfect. But it should be worthy of attention. Memorable, even evocative in the way it triggered emotion. The kind of product that prompts someone to say, “I didn’t know I needed this, but now I can’t live without it.” You can refine mechanics over time, but the emotional experience of a first interaction is singular and unrecoverable. Don’t settle for viability. Build something worth remembering… and make it awesome! ❤️🔥
26 Comments -
Benjamin Flores
The Uber founder thinks we're heading toward "LeBron James-level plumbers." Travis Kalanick argues that as we race to superintelligence, the bottleneck for most projects becomes skilled manual labor. Think about it. Every non-manual task gets 10x faster. Strategy, code, design, reports, all done in minutes. The bottleneck shifts to moving atoms. Building the office. Wiring the servers. Installing the equipment. Fixing the pipes. Skilled manual workers become the most in-demand people in every company. Not because blue collar is all that matters. But because it's now the only bottleneck to growth. When every company can spin up software in hours, the constraint is who can build the physical thing. Demand for skilled trades goes through the roof. This take is particularly interesting coming from the Uber founder. Everyone told him the taxi market wasn't big enough to build a billion-dollar company. He proved that supply creates its own demand. When someone has that kind of understanding of non-intuitive supply and demand dynamics, and he tells you skilled labor is the next bottleneck, you listen.
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content