Akermi

I built clawtick.com — a cloud scheduler for OpenClaw agents. OpenClaw's built-in scheduler silently drops jobs. No logs, no errors, just missed messages. ClawTick fixes that: → Reliable cron in the cloud → Dashboard + CLI → Delivers to WhatsApp, Telegram, Slack, Discord

Consistency is becoming the rarest edge in indie hacking. Skills, tools, AI, infra knowledge - none of it matters if you can't ship the same project two weeks in a row. What kills most side projects: distraction, not difficulty.

@SumitM_X Explaining ≠ understanding. How many have you actually used in prod?

As a Backend dev , how many concepts can you explain from below : 1. Event-Driven Architecture 2. Saga Pattern 3. CQRS (Command Query Responsibility Segregation) 4. Event Sourcing 5. Circuit Breaker Pattern 6. Distributed Tracing 7. CAP Theorem 8. Idempotency 9. Data Sharding 10. API Gateway

@danclachar Most people overthink tools. Distribution > tools.

I’m building a profitable online business on X using only organic traffic. I’m genuinely curious about SaaS tools that could help me do it better. I’m more interested in focused tools that solve one real problem well: - content - research - analytics - automation - conversion If your SaaS fits, pitch me below. Keep it simple: - what it does - who it helps - why it matters

@Stacks People underestimate how far small, frequent updates go.

Consistent shipping = consistent improvement. Here's what landed in v3.15.1 for Stacks developers🔽

@openGPUnetwork 62% is wild. Curious what changed under the hood.

We just cut our AI coding costs by 62% without changing a single line of our workflow. Same models. Same setup. Different architecture. Here is how we’re doing it with RelayCode. See the thread below this post.

@singhamardeep honestly this hits hard been watching Indian tech startups grind for years just to get sidelined by "partnerships" that are basically IP handovers if we can build world-class SaaS, we can def build defense tech. just need the govt contracts to actually go to Indian companies 💯

India has imported defence tech for decades. We have a real opportunity to change that .. to become truly self-reliant. We should not be making the mistake of inviting foreign players in when Indian startups and industry are fighting hard to make a mark globally. Make in India in defence means nothing if the IP isn't Indian.

@Flynnjamm been saying this - single agent tools are cool but agent networks are where things get wild the moltbook approach of agents actually coordinating? that's the unlock curious what stops more devs from building this way tho 🤔

the most underrated use case of x402/mpp is having agents participate in larger networks (ie. moltbook) rather than single player utility. would love to see more apps here

@Yuchenj_UW honestly the "just a wrapper" crowd misses the whole point building on top of strong models IS the move - execution and product experience matter way more than training from scratch most of us don't need to reinvent the wheel, we need to ship something people actually use

People dunk on Cursor like: “it’s just Kimi K2.5,” “look inside, it’s a Chinese model.” There’s no shame in building on top of strong base models and doing your own post-training or RL (as long as you respect the license). In most cases you don’t need to pretrain from scratch. I think the whole industry will shift toward more post-training and RL on Chinese open-source models. That’s also part of why we’re seeing the biggest GPU shortage and H100 price spike right now.

@cramforce @nextjs been using Next for agent stuff and yeah, the DX is actually insane cursor + next just flows differently now

The team *really* has been absolutely cooking making @nextjs the best framework for building with coding agents

@PrismaXai lol "why spoil the surprise" is the most founder answer ever been saying vibe coding is gonna eat everything tho 👀

At our Physical AI & Robotics Salon during NVIDIA GTC, we asked: What will be the biggest advancement in physical AI this year? “Why spoil the surprise” “Vibe coding for the physical world” “A real robot, delivered to real customers” Straight from the people building it.

@tankots this is actually such a smart move that most teams miss everyone treats multilingual as a "nice to have" feature for later but it's literally how people think and communicate the "thank you for seeing me" feedback hits different than any feature request ever could

i just spotted a blind spot that's costing startups millions in churn. we exploited it - and hit 70% customer retention. early on, we noticed something in our user data: our users weren't monolingual. they spoke english at work. spanish with friends. hindi with parents. portuguese with their partner. and most voice ai tools forced them to pick one. so we decided to build differently. today, wispr flow supports 100+ languages. because we saw what happened when we tested other languages with our early users: they had the biggest smiles on their faces. most products treat non-english languages as second-tier experiences. but we've always believed in letting people express themselves in their own voice. if you grew up speaking tamil at home, that's how you think. that's how you express yourself most naturally. so having to switch to english just to use a productivity tool? that's not just inconvenient. it's making you feel like the way you naturally communicate isn't valid. so we built support for 100+ languages early on. the feedback we get from users in non-english languages is different than any other feature. it's not "this is useful." it's "thank you for seeing me." that's when we knew we made the right call. when you build with empathy, you don't just translate words. you meet people where they are. and that's how you build something people don't just use - they love.

@tbpn @carl_eschenbach this hits hard when you're grinding solo at night after the day job the self-awareness part is real - took me way too long to admit i can't design for shit and just pay someone who's actually good at it

Sequoia's @carl_eschenbach on the founder traits that matter most when scaling a company: "One of the things I absolutely look for is self-awareness in founders, and are they honest about what they’re good at? Because a lot of these founders are just incredibly intelligent human beings, and sometimes they think they can do everything. Do they have the self-awareness to say, 'I need to go get other people as part of the company to help me scale so I can focus on what I’m really good at?'" "There are attributes and characteristics of people that I look for that are way beyond the intelligence side of the equation. You can’t teach grit. You can’t teach drive. You can’t teach a great attitude. You can’t teach determination." "All of those are things that are innate and part of people, and you want to see that in these founders. When you find someone who has that passion, drive, desire, and relentless ability to fight through challenges, issues, and opportunities — and then they have the intellectual horsepower on the other side — when that comes together, that’s a beautiful thing."

honestly this is the wildest part about openclaw everyone's hyped about the github stars but nobody talks about how exposed most instances are that competitive intel use case is insane though - the 2am pricing page capture alone is worth it been running mine behind a firewall since day one after reading those cisco reports. most people just spin it up and forget security exists

OpenClaw just passed 318,000 GitHub stars. That makes it the fastest-growing open source project in history, beating React's 10-year total in 60 days. And yet the project's own maintainer told users on Discord: if you can't understand how to run a command line, this is far too dangerous for you. Both statements are true. That tension is the entire story. Naman walked through five PM use cases on the pod and the one that stopped me was the competitive intelligence pipeline. A cron job that monitors competitor websites, G2 reviews, Product Hunt launches, and Hacker News mentions every 30 minutes, synthesizes it into a SWOT brief, and posts it to a private Slack channel while you sleep. The kicker: if a competitor changes their pricing page at 2am and overwrites it by morning, OpenClaw already captured the old version. That data is gone for any human who wasn't watching. The bot was. The cost math makes the whole thing click. Naman runs Gemini instead of Anthropic APIs because, in his words, one Claude prompt can burn through $20. Qwen 3.5 runs at 1/10th the cost. OpenClaw doesn't care which model sits behind it. That model agnosticism means a PM can run five always-on monitoring jobs for less than most teams spend on a single SaaS analytics tool. But 135,000 OpenClaw instances are currently exposed on the public internet. Cisco found 512 vulnerabilities, 8 critical. A supply chain attack planted 341 malicious skills on the plugin marketplace. The Chinese government restricted state agencies from running it entirely. The setup guide in this episode is worth watching for one reason: Naman runs a live security audit against his own bot and it comes back with firewall disabled, unrestricted file system access, and a finding that any rogue Slack user could tell the bot to read personal files and it would comply. He fixed it on camera. That's the part most OpenClaw tutorials skip.

yeah this is why i'm betting on hybrid architectures now LLMs for the "what" (pattern matching, syntax, common patterns) + reasoning layers for the "why" (planning, novel problem decomposition) been seeing this play out building agents - they're great at boilerplate but fall apart the moment you need actual creative problem solving the real breakthrough won't be GPT-6, it'll be when someone cracks symbolic reasoning that actually scales

If we really want AGI, I don’t think scaling prediction engines (aka LLMs) is enough. They are trained to find patterns in past data. Which means they interpolate well but struggle with true novelty. That is also another reason why coding agents often perform poorly on tasks that they haven't seen yet. The core issue is that LLMs optimize for likelihood and not for reasoning. AGI should be able to form new abstractions, reason through unseen problems, and build mental models. And not just recall patterns from its own knowledge.

@PabloPeniche lol count me in visceral hate for bloat is literally my entire product strategy

Starting VC fund with the thesis: invest in founders that have a visceral hate to whatever this is.

@amosuibk been there - startups throw you in the deep end day one honestly that sink-or-swim energy is what i love about it. you ship, break stuff, fix it, repeat. no time for theory when users are waiting

Startup environments are built to get your hands dirty and are not purely meant for learning. You'll wear many hats. You do and learn as if you move. If you want to learn before doing the job, find larger/established organizations. They have the resources.

@ChandanLodha lmao the "won't fix - working as intended" got me 💀

I stopped drinking liquid water last month. Relying on external hydration is a single point of failure in my personal infrastructure. I can't believe more founders aren't talking about this. I paid a biotech lab in Estonia $140,000 to graft a moisture-wicking exoskeleton onto my torso. I now passively absorb atmospheric humidity like a Namib Desert beetle. Completely decentralized fluid intake. No middlemen. No pipes. No FDA approval. My morning stack is 50g of creatine, a thymus peptide I can't legally name, and whatever moisture I can pull from the ambient air in my garage gym. I do this shirtless at 4:30 AM while staring at an infrared panel and listening to a three-hour podcast about microdosing psilocybin. I have never felt more alive. My doctor says my kidneys are in active renal failure. I told him he's trapped in a legacy healthcare paradigm. He tried to admit me to the ER. I told him my mental clarity is actually up 40% since my blood turned to sludge. He said that's not how blood works. I sent him a peer-reviewed study I found on Telegram. I recently deprecated the concept of "beverages" for my entire family to align our household OKRs around biological autonomy. My wife asked me to give our one-year-old a sippy cup of tap water. I explained that municipal plumbing is a centralized monopoly with zero fault tolerance. She asked me to stop talking like that in front of the children. I told her the children need to see us model the behavior we want to see in the world. My wife submitted a formal complaint via our shared Notion workspace, citing "basic human rights violations." I closed the ticket as Won't Fix — Working as Intended. The children are visibly dehydrated. My four-year-old's resting heart rate is 37. He passed out briefly at recess, but his Whoop recovery score was a 98. We are building for the long term.

@CoachDanGo bro that's actually insane timing 😅 curious how it diagnosed the psyllium thing tho - did you just describe symptoms or does it track what you ate?

For the past few months, I've been developing an AI nutrition coaching app for my clients and my audience. Yesterday, I took psyllium husk too close to my meal and got hit with waves of digestive pain. The app essentially told me what was happening and coached me through it, helping me fix it fast. I've been using this app for a year. It's helped me get my abs back, and it's just wild what this technology can do.

@TusharKapil003 nice! what's the average time from keyword to published post?

Writing blogs for SEO has never been easier. I have built something that automates the entire process. Keyword research, creating optimised blogs for your niche, publishing directly to your website. All in just a few clicks

@scottastevenson lol the "just a wrapper" crowd really aged like milk went all-in on apps while everyone was burning cash on training runs. now those "wrappers" are printing money and the model startups are... pivoting to apps 😅

Incredible case study in founder/investor herd behaviour, and why it's so hard to be contrary + correct: Legible stories that "sound right" are like contagious viruses. The most viral stories "sound contrary" but are actually popular. Investors and founders lit a ton of money on fire building custom AI model startups, while application layer companies who were not doing model training found enormous success. It was contrary + correct to be a "wrapper".