Aditya Chichani

Meta • 2K followers

We’re ~3 weeks away from the submission deadline for SIGIR eCom’26. 📅 April 27 (11:59 PM AoE) If you’re working on search, recommendations, or at the intersection of IR + LLMs + UX - we’d love to see your ideas. This year’s theme, Agentic-driven user interaction & experience, is especially aligned with what many of us are building right now - moving beyond static ranking systems toward interactive, adaptive experiences. In addition to papers, there are two exciting ways to participate this year: 🔹 Data Challenge In collaboration with the TREC Product Search Track, focused on query reformulation for task-oriented eCommerce queries. 🔹 Agents Anonymous An interactive, “group therapy”-style session: submit a 1-minute problem pitch on a real-world challenge you’re actively facing, and get feedback from an expert panel + peers. 📍 Melbourne (in-person, July 24) 📄 CFP: https://lnkd.in/grUbQMxi 📥 Submissions: https://lnkd.in/gYES-NUD → “2026 SIGIR Workshop on eCommerce” Dean Alvarez, Surya Kallumadi, Yubin Kim, Tracy Holloway King, Andrew Trotman

Joseph Zhang

CatchBack Cards • 3K followers

One underrated part of the YC W26 batch has been getting exposed to really great products built by other founders. A few companies whose tools I’ve genuinely loved using recently (I wasn't sponsored or promoted by any of these companies). Vercel Deployments, hosting, envs, previews — all the stuff I never want to think about again. It just works, and startup credits/accelerator support makes it easy to adopt early. Cursor The IDE experience is a game changer. Prompting directly in-editor, accepting changes, and easily reverting/rolling back makes iteration insanely fast. It’s become central to how I code and iterate. Anthropic Even with similar underlying models, Claude consistently gives better answers for my use cases. I’ll admit it — I hate living in the terminal (fake coder, I know), so prompting in a sidebar fits my workflow way better. Sometimes it hangs, but improvement on the frontend feels very real. SideShift, not YC, but necessary for B2C I have a love and hate relationship with SideShift because I love seeing the metrics on UGC creators, but finding good ones is a pain in the ass. Sophia and I spend way too much time sifting through applicants who aren't really worth the time to invest in. Supabase (YC S20) Auth, storage, database, RLS — all in one place and actually pleasant to use. Spinning things up and shipping is fast. RLS policies alone saved me from a lot of pain compared to other stacks I’ve used. YC credits/benefits helped early scaling. PostHog (YC W20) Product analytics that actually help you understand user behavior: event tracking, funnels, feature usage, session replays. I didn't know what PostHog was before starting YC, and now I can't imagine building without it. Mintlify — (YC W22) Docs that actually look good. Clean UI, fast setup, and it makes writing and maintaining documentation way less painful. I enjoy prompting AI to help flush out my language. ElevenLabs Insanely realistic voice generation. We’re planning to use it heavily for ad creatives and rapid iteration on audio without needing studio time, and YC startup credits make experimentation affordable. Brex (YC W17) Startup banking and cards done right. Literal best customer service (above Amex plat and Capital One Venture X concierge.) Would gladly park next round's fund in here with these guys as well. Also, the DashPass that comes with the bank account feels absolutely so amazing. The feeling of getting a BOGO deal and it being on Brex, amazing. Clerky (YC S11) Incorporation, equity, and legal docs without the usual headache. I still have to redeem my Clerky refund, oops, but I'll get that credit back when I have the time to sit down and redeem, haha. Linear I worked at McKinsey touching Jira boards before and wanted to die. Linear is just miles easier to deal with. My DMs are open to founders building tooling! We're building a platform to revolutionize the collectibles and trading card game industry and love trying things out.

Tario You

67/Labs • 7K followers

today a YC founder asked me how we approach hiring for OpenTrade (YC s26), and to be honest I still have zero clue. Like just look at our museum of RIPs. Dylan Pak overpaid for a UGC campaign manager by 3k that didn't even cook. Ojas Kandhare immediately regretted work trialing an engineer who seemed too good on paper, which he was. and I had to personally let go 2 interns because they preferred larping over doing any real work. but as they say, what doesn’t kill you makes you stronger. so here are 3 ground rules we now set for every hiring decision at OpenTrade (YC s26)     1.    Do take home deliverables. Dylan Pak lives by them as he says you can easily tell if a marketer is worth their weight within 24 hrs by the way they approach doomscrolling and how they put that on a creative brief. But the main thing your looking for on interviews is whether you learn new things from candidates and they have a history of thinking deeply about problems.     2.    Work trial, work trial, work trial If they somehow pass step 1, do a 1-2 day paid work trial. Main thing here is your seeing if they can actually do a dent in tackling their assigned problem. Because if they can’t, just don’t hire them. As our CBO says it costed him 100x in his time to train one of our interns to Cluely marketer caliber at that stage than to just pay someone with the same hunger to learn, but who cooks right away.     3.    Never settle for B players And the difference between A and B is how they view the same problems. For example, instead of staring at problems like B players, A players actually take full-responsibility in solving them. Because when your moving at warp speed, you can’t settle for micromanaging. Or it just slows the entire startup down. TLDR; this is our personal take on hiring, and a bit of sam altman sauce I read last night. but if your mighty talented, hit me up, or if you know any good referrals, we also pay you $6 or 7k if we hire them :)

Henry Shi

Anthropic • 81K followers

I stepped back from my $200M+ rev/year startup and joined Anthropic to help build safe AGI. Here's my resignation letter and my 9-step transition process so my company could continue to grow 50%+ YoY after my exit: I’m sharing this openly as there are many guides on starting companies, but none on resigning and transitioning well. As AI continues to accelerate and consume SaaS, many unicorn founders have reached out asking about transitioning best practices. They don't know how to set the company up for continued success and leave while preserving the legacy they've built. So here's what I learned after spending 9 months preparing my exit: Step 1: Write down your story before you tell anyone I spent 3 weeks documenting my purpose, mission and what I accomplished during my tenure. That document became my north star Step 2: Talk to your co-founder privately That Christmas call with Hussein was one of the hardest conversations I've ever had. Make it an honest heart-to-heart conversation. Share openly and don't hold back. Align on the successor and the timeline before anyone else knows. Step 3: Talk to board members 1:1 and not in a group In group settings, board members often perform for one another, whereas individually, they are honest and collaborative. I spent months on these conversations, so by the official board meeting, every member was aligned. Step 4: Hold the official board meeting This should happen after your successor is finalized with a clear timeline and next steps. Step 5: Email investors with your narrative and legacy attached Share a complete narrative, specific accomplishments with numbers, and a succession plan. Own your story. Step 6: Talk to C-suite and key employees 1:1 before the all-hands I scheduled 30-minute 1:1s with 15+ people over two days. It was exhausting, but the most meaningful conversations of the transition. Step 7: Make the company-wide announcement Then hold office hours and show up for the questions. Step 8: Use the window before your last day You can't get fired, and this is the only time you have full founder authority with zero consequences. Push bold initiatives, set up easy wins for your successor's first 90 days, and document everything. Step 9: Make the public post count My LinkedIn post got 100k+ impressions, hundreds of comments, and dozens of opportunity DMs. I told the story with specific numbers, celebrated the team by name, and ended with a clear CTA. Remember that your transition only succeeds if the company thrives after you leave. If you are considering a transition to an AI lab or just thinking of an exit, start planning now instead of waiting until the last moment. Subscribe to my Substack for the complete resignation letter, the detailed 9-step process with specific timelines for each phase, the conversations I had with my co-founder, board members, and investors, and every hard-earned lesson from my exit. 👇 Link in first comment