@sachinrekhi@sachinrekhi I recommend you try the Obsidian web clipper. I have been using @obsdmd as my knowledge base paired with Claude Code - everything from my HypeNote meeting transcripts, Perplexity research all go into my Obsidian vault into project specific folders.
Claude code competitor research hack:
1. Ask claude to find competitor Canny boards
2. Get claude chrome extension to scrape all feature requests, bugs, pain points and save in a md file
3. Use the md file for all SEO content with Claude code-
Claude code will create pages that speak directly to the pain points of your ICP
Content that will -
- Rank better on Google
- Get cited by AI engines
- Turn leads into customers
An NVIDIA powered farming machine uses Al vision and precision lasers to eliminate weeds in milliseconds without herbicides and without harming crops, a potential shift toward chemical free agriculture
The PMs who win in the next wave won't be the ones who figured out how to prompt to build. They'll be the ones who figured out how to run 10x the customer learning with the same team.
Here's why that matters right now.
AI has handed engineering teams a jetpack. Cursor. Codex CLI. Claude Code. The delivery side of product development — build, specify, launch — is being automated at a breathtaking pace. But as @AndrewYNg recently pointed out, the real bottleneck today isn't coding. It's discovery.
While everyone raced to accelerate shipping, the question mark moved upstream. We now have the ability to build faster than we've ever been able to learn. And building fast on the wrong insight isn't speed — it's just expensive mistakes, sooner.
The good news: the same AI revolution is quietly making discovery dramatically more powerful too. A few of the emerging use cases:
1️⃣ Analyzing feedback at scale. What used to require a researcher and two weeks can now be done by a PM in an afternoon — feeding thousands of NPS verbatims, support tickets, or app reviews into an AI and getting back a structured synthesis of themes, patterns, and verbatim quotes.
2️⃣ Automating feedback rivers. Tools like Reforge Insights, Enterpret, and Kraftful now continuously monitor customer feedback across every channel and surface actionable signals without anyone having to manually triage.
3️⃣ AI-moderated user interviews. Platforms like Reforge and Listen Labs are making it possible to run interviews at a scale that was never feasible with human moderators — turning what used to be 10 interviews into 100.
4️⃣ Discovery via prototypes. With vibe-coding tools like Lovable, v0, and Bolt, PMs can now build functional prototypes and gather real behavioral data — heatmaps, drop-offs, in-product surveys — before a single line of production code is written.
5️⃣ Natural language metric analysis. Ask your database a plain-English question, get a chart back. No SQL. No waiting for a data analyst. The feedback loop between a hypothesis and an answer just collapsed from days to minutes.
The teams that wire these workflows together won't just be better informed. They'll develop a sharper product intuition — the kind that David Lieb (Founder of Google Photos, Partner at YC) described as "the world's most sophisticated machine learning model ever created."
Join me Thursday, March 5th at the Lean Product Meetup with @danolsen in Mountain View, CA where I'll be sharing the exact 10 AI discovery workflows I now rely on to help me decide what's worth building faster 👉 https://t.co/mUpRvvjmD1
We just passed React on GitHub stars. 🦞
Let that sink in. A personal AI assistant built by a lobster-obsessed Austrian and an army of crustacean enthusiasts just outstarred the library that powers half the internet.
We shipped 90+ changes today. They shipped a conference.
Productivity gains have accelerated with every new wave of technology - each one riding on the path paved by the previous wave from electricity >> telecommunication >> internet >> social/mobile >> now AI.
Productivity gains from AI would be realized sooner than we expect.
AI models made you a superhuman.
For 20 years, software barely moved. Now a guy in his bedroom with AI can compete with a 10-person company shipping slow and overcharging for bloated UI.
You don’t need to reinvent anything. Just take a boring cash-cow product and make it 10x better:
•Fair pricing
•Cleaner design
•Faster onboarding
There’s a gap. Your only job is to close it.
@marclou@steipete 💯 OpenClaw has brought a seismic shift in what is possible with AI. People are thinking beyond text generation to e2e workflows and task completion.
@andrewchen@andrewchen its not the frequency of use - quality matters.
Most people are using AI to generate content and fix text while some have evolved to treating it as a work partner offloading critical work. That shift in mindset is where the real divide is.
I haven't felt like this since 1989.
I've had a computer since I was 8. Back then, it was BBSs, IRC, Prodigy, AOL. Small communities of people tinkering and building together. You know that feeling when you can't stop because you're learning too fast? I haven't felt it in decades.
Last week, it came back.
I've been using OpenClaw for real work beyond personal productivity, and I haven't shared anything publicly because I'm learning so much by getting my hands dirty. Multiple nights up till 2am doing DevOps work I haven't touched in 30 years. There was no deadline. No customer waiting. I just couldn't stop.
When I mentioned this yesterday, @awwstn captured it perfectly: "hiten out here proving it is possible to use openclaw without posting incessantly about it."
I'd been thinking about posting a worklog or journal. Document the journey. Share the wins. But every time I started to write something, I stopped. There's too much attention being put on simply getting attention itself versus figuring out how this stuff can truly help us with tasks we couldn't dream of not having humans do.
Too much opportunity out there to waste time on documentation theater.
Don't get me wrong. I'll share. That's what this is. But I wanted to understand it first. I wanted to break it, fix it, and deploy it across multiple Slack workspaces. Give it a GitHub account and watch colleagues use it in ways I didn't expect.
Everyone's talking fast. You learn faster when you're quiet.
I set up a bot called Bob. He has his own GitHub repo. His own email address via AgentMail. I spun up multiple bots across different Slack workspaces and companies. I introduced the bots to colleagues and friends. Some of them started building immediately.
One friend shared their own non-OpenClaw agent's repo with Bob and had Bob start coding solutions to their agent's workflow problems.
Personal productivity tools save you time. This gives you capabilities you never had.
The first few nights were rough. I was deep in the terminal, setting up a VPS, debugging integrations, dealing with errors I hadn't seen in decades. The system used to be called Claudebot, then Moltbot (because trademarks), and now OpenClaw. The naming evolution alone tells you how early and chaotic this is.
But by the end of the week, I could spin up a fresh instance in 15 minutes and add it to Slack.
That's the journey. From 2am terminal sessions to frictionless deployment. That compression matters.
I'm not selling you on a tool. Something fundamental shifted for me last week. The AI model didn't do it. Neither did the integrations or the DevOps.
The shift happened when time and capability stopped limiting me. Now imagination is the constraint.
The work teaches you things the discourse never will.
There's more to share. Bob and the other bots are doing interesting work. I'm learning what happens when you treat AI as a colleague instead of a tool, and what shifts when non-engineers get capabilities that used to require engineering teams.
But first, I had to build it.
WELCOME TO THE ROOM: Satya Nadella's Lesson in Executive Accountability
Most people treat a senior promotion as a destination; in reality, it is an invitation to a higher level of pressure where excuses are considered a form of professional failure. In "The Room," the distance between a "theory of success" and "actual success" is measured by one thing: Intellectual Honesty.
If you are waiting for more resources, more time, or more favorable conditions to win, you aren't leading—you’re whining. True leadership is the act of "manufacturing success" within the constraints of reality. It requires the scientific rigor to align your resources to your theory, the telemetry to admit when that theory is failing, and the courage to pivot before the runway disappears.
The Bottom Line: You are either a generator of clarity or a creator of confusion. If your "dots don't connect" from your current headcount to the final result, you are just managing decline. Stop talking, get the telemetry, and operationalize the win. Anything else is just noise.
https://t.co/T3JQrXRrDu
@satyanadella 📢 Copilot search, a powerful AI powered enterprise search with hundreds of connectors, is coming to Microsoft 365 Copilot customers! #EnterpriseSearch#Copilot