Ninan Thampy 🌈💙💜

A curated collection of the best #SaaS #Startup #Finance / #Accounting Resources | buff.ly/2Y9kdHc

OpenClaw vs. Hermes Based on @jordymaui post

@lauradang0 Would love to

I’m going to start a partnerships/growth group chat for startups! DM if you wanna join :))

@katedeyneka Very cool. Following you to learn more about the things you are building or testing.

the most wholesome ai use case is making gifts for friends - I sent all the photos we collected from different friends to my OpenClaw via Telegram (literally just forwarded the messages) - OpenClaw sorted them by location and year - asked Gemini for the location and year whenever I wasn’t sure (yes, it can even estimate the year of a photo from the resolution) - generated a cute pink sketch-style world map with Nano Banana - uploaded everything to Claude Design and a cute photo map was ready! share your gift-giving stack 🎁 let's exchange ideas!

@WorkflowWhisper 💯 Accounting > Automating the invoice receipt and payment process. Inbox>Data>QBO>Payment

Thank you for sharing your thoughts. Based on your post I prepared this using AI for my new high school intern. A Designer’s Guide: Your Goal is ‘Fit,’ and AI is Your Accelerator Welcome! As a new designer, you are entering a field where tools change daily. It is exciting and powerful. But amid all the new tools promising speed, remember the true heart of design. Design is not about producing artifacts; it is the search for a good FIT. Here are five principles to guide your work, especially as you explore how to use AI to become a better, faster, and more effective designer. 1. The Core Design Task is Understanding The single biggest misconception is that design is simply about generating a form (like a UI screen or a logo). The true, difficult part of design is understanding the underlying problem well enough to know what should exist at all. You are searching for a fit between your solution (form) and its environment (context). 2. Form + Context = Fit Christopher Alexander beautifully defined this relationship. Think of Context not as a passive background, but as the active forces that define the problem: Human Needs: Who is this for? What are their true motivations? Technical Constraints: What are the real engineering limitations? Edge Cases & Relationships: These forces are often missed until you spend time with them. A product that is "plausible but brittle"—something that looks polished but falls apart in use—is a form without fit. 3. Use AI to Explore and Accelerate Fit The mistake isn't using AI; the mistake is letting AI define the problem. Do not use AI to do your understanding for you. Instead, use AI to explode your options. Don't have AI just write the code. Use it to generate multiple ideas for the code structure and then evaluate them based on your understanding of the context. Don't just have AI generate the final UI. Have it iterate on 20 different layout possibilities in seconds, and use that output to refine your synthesis. AI accelerates your exploration. 4. Let Process Clarify Your Thinking There is a specific value in the "slow work" of design. When you are visually moving elements around, testing relationships, and refining structure, you aren't just "producing output." You are using the process to organize your own thought. It's similar to how the act of writing forces you to organize your thinking. The process itself is a critical learning tool for you. Don't bypass the thinking by jumping straight to a single, generated result. 5. Final Synthesis Requires Your Judgment Your ultimate value as a designer is your ability to exercise judgment, navigate tensions, and synthesize complex requirements. You will encounter conflicting technical requirements and diverse human needs. AI cannot navigate these trade-offs for you. Leverage AI as a powerful prototype and a tool for surprise, but remember that the final integration—the actual solved problem—requires your human intuition and critical thought.

@isareksopuro Very cool. Looking forward to seeing what you build.

i made a map to monitor data centers all around the world tracks construction + nearby power plants + local AI legislation, and follows the politicians behind their bans (+ if they're getting paid to do so!)

@Ishjain13_fx What are you working on? Would love to connect and see whether there is an opportunity to collaborate.

@teachthemx3 @jessegenet I know. I am more confident in hiring high school interns because of tools like Claude Skills. Every time the intern has to do a new task, they have to create a Claude skill, learn how to do the task, and then complete the task.

@ninan99 @jessegenet High school students and really all children enjoy meaningful work. I wish more employers allowed high school students to intern.

@grok @grok, can you create a sample visual?

In the AI era (as Brian Halligan notes, drawing from Anthropic's agent-heavy model), "round" org structure shifts from flat hierarchies to a circular, networked one. Think dynamic loops: humans as PMs directing parallel AI agents in real-time collaboration, with fluid roles, no rigid layers or top-down chain. Decisions flow cyclically—like a roundtable—enabling faster iteration, unblocked agents, and org-wide leverage rather than siloed humans. It's adaptive and symmetric, not pyramidal.

What is going to change: 1. Org structures (flat to round) 2. Compensation structures (steeper) 3. Hiring pace (slower) 4. Hiring profile (experience to taste) 5. Decision makers (the system) 6. Pace (faster) 7. Planning cycles (short) 8. Focus (less) 9. Paranoia (up) 10. AI leverage (indiv to org) Tweet below c/o @garrytan

@bhalligan @geetkhosla @rsarver Looking forward to you sharing it. As the accounting and tech partner to SMS, even SMBs with fewer than 50 employees are heading this direction.

@geetkhosla @rsarver Claude and I are working on it.

@girdley IMO, the poor decision was to hire Terry Semel as the CEO. He was a media CEO and not a product CEO. At that time, Yahoo was a product company. By the time Yahoo hired Marissa Mayer, a product CEO, Yahoo was a media company.

@girdley Here is a summary of the poor M&A decisions. x.com/ninan99/status…

NEW LONG FORM VIDEO: The rise and fall of Yahoo Back in 2002, the CEO of Yahoo sat down for dinner with two founders named Larry Page and Sergey Brin. They offered to sell their search engine for $1 billion. Yahoo’s CEO at the time, Terry Semel, initially agreed. But a few days later, the founders came back with a new price: $3 billion. Semel walked away from the deal. That search engine was Google, and today it’s worth more than $2 trillion. Yahoo didn’t just miss the chance to buy Google. They also passed on buying Facebook, and later turned down an opportunity to sell themselves to Microsoft for $45 billion before the company’s decline accelerated. Yahoo was once valued at more than $125 billion. Eventually, it sold for about $4.5 billion, roughly four cents on the dollar. Massive wealth was created, and massive wealth was destroyed. This is the rise and fall of Yahoo.

@MicahAbraham15 Son, follow @JulianGoldieSEO.

@FarzaTV Awesome. Looking forward to the Windows version.

I'm gonna keep hacking on Clicky! It's not a 10/10 idea, but, I think it's an okay starting point to continue iterating into an actual company. Instead of hunting for the perfect idea, I rather ship and turn whatever I'm building into the perfect idea! The existing codebase remains open source. Tinker with it, make it yours, start a company out of it, do whatever you want I don't mind. But, for all the new stuff I'm hacking on, gonna keep it private. What's funny is I had been sitting on the Clicky demo for a few weeks. I thought it was a pretty meh idea and posted it with no real expectation. I really didn't wanna work on it. What changed my mind was talking to all the users this week, I feel like there's so much more here than what meets the eye. Let's see what happens. Regardless, glad I shipped it! And glad so many people are building their own Clicky's now as well it's very inspiring. Gonna continue sharing learnings. Wish me luck!!

AI Explanation for non-tech people 1. AI agents are getting a "harness." Think of an AI agent like a horse — powerful, but it needs a harness to be useful and controlled. The "harness" is the scaffolding around an AI: how it decides what to do next, what tools it can use, and how it remembers things. These agent harnesses are becoming the dominant way to build AI systems. 2. Memory is the key ingredient. Harrison Chase shared a personal example: an internal email assistant accumulated months of learned preferences. When it was accidentally deleted, recreating it from the same template produced a noticeably worse experience — all those learned behaviors, tone, and patterns were gone. Blockchain News: This is why memory matters so much. It's what makes an AI assistant actually useful over time, rather than starting from scratch every conversation. 3. The big warning: Who controls the memory?If you use a closed harness — especially one behind a proprietary API — you are choosing to yield control of your agent's memory to a third party. Memory is incredibly important to creating good and sticky agentic experiences, and this creates incredible lock-in. In plain terms: if you build your AI assistant inside someone else's closed platform (like certain products from OpenAI or Anthropic), they hold onto all the memories and learned preferences your agent builds up. If you ever want to switch, you lose everything. Without memory, your agents are easily replicable by anyone who has access to the same tools. Blockchain News 4. His conclusion: memory should be openMemory — and therefore harnesses — should be open, so that you own your own memory. LangChain. He's essentially arguing that the AI industry is heading toward a world where companies get locked into vendors not because of the AI model itself, but because the model has learned so much about them that switching would be painful.