J Chaudhary

@RoundtableSpace Is anybody thinking about prompt injection and security? What about context overloading?

THIS TOOL SCANS YOUR PROJECT AND INSTALLS AGENT SKILLS BASED ON PROJECT TECH STACK github.com/midudev/autosk…

Today we're excited to announce NO_FLICKER mode for Claude Code in the terminal It uses an experimental new renderer that we're excited about. The renderer is early and has tradeoffs, but already we've found that most internal users prefer it over the old renderer. It also supports mouse events (yes, in a terminal). Try it: CLAUDE_CODE_NO_FLICKER=1 claude

You can now set the permission mode for coding tasks in Dispatch. We recommend using Auto mode for the safest and most seamless Dispatch experience but any of your allowed permissions are available. Note if you use Bypass Permissions you need to approve session start.

@RoundtableSpace Looks cool but as long as users can not own this locally the problem is not solved for many businesses

GENERAL PURPOSE AI IS STARTING TO GIVE WAY TO CUSTOM MODELS BUILT FOR SPECIFIC WORK. Oumi is pitching a faster path to build your own model in hours instead of months.

@lukepierceops Yeah and there are people like us who have been building agency from last 10 years to reach 7 figures and wondering are we left behind?

There is a guy on here with 35k+ followers Openly admits he's been running an ai automation agency for LESS THAN A YEAR His LinkedIn bio claims he's "scaled my AI agency to near 7-fig months" On LinkedIn, his community starts the same time he started is ai automation agency (11 months ago) Nobody... I mean nobody, even the TOP operators I know do not scale to 1M months in less than a year Literally claiming to have built a $10M business in 11 months... But is selling a $2,000/yr community LMAO And all that as a mid 20 year old with no experience. But it gets better... You go to his agency website, and all of the testimonials are fake names and fake companies. These are the people I am going to WIPE THE FLOOR CLEAN with The level of LARP is UNPRECEDENTED I WILL BRING THE PEOPLE THE CORRECT INFORMATION AND MAKE SURE THESE SCAMMERS GO TO ZERO

@tom_doerr LI automations are hard. Thanks for sharing

Self-hosted AI LinkedIn automation tool github.com/eracle/OpenOut…

@arvidkahl I have collected quite a few in google sheets

The discovery of Claude skills still is an unsolved problem. I learn of the most insane high-impact skills from throwaway mentions in some reply somewhere. My mobile browser is full of links like these. What’s the most impressive Claude skill you’ve ever installed?

@karankendre We built a voice agent for couple Of client ls last year took 3 weeks. Today same thing takes 3 days. The cost curve is brutal for legacy call centers. The ones who survive will be the ones who own the agent layer.

Call Centers Will Go Bankrupt If People Start Building Their Own Voice AI Agents The only thing you need to do to make thousands of dollars is: 1. Partner with salons, clinics, and other service-based businesses 2. Deploy a custom AI agent to handle FAQs and appointment bookings 3. Price it at $1000/m roughly one-third (or less) of what a human receptionist actually costs a business This article is a complete playbook to do exactly that

@NickSpisak_ Most AI agent failures are silent. The agent completes the task, returns a confident response, and logs a success.. but the real-world outcome is wrong. Testing for observable behavior is hard as well and that langchain piece is interesting

Most AI agents break in ways you don't notice until it costs you. LangChain just open sourced how they test theirs - and the approach is worth understanding even if you never write code. Here's the key takeaways in plain English: → Don't test everything. Test the specific things that matter most. → Getting the right answer isn't enough. Getting it fast and efficiently is what separates good agents from expensive ones. → Every time your agent makes a mistake, turn that mistake into a test so it never happens again. → Tag your tests by category so you can run just the ones you need instead of waiting for all of them. Why this matters for anyone building with AI: → Your agents get more reliable over time instead of randomly breaking → You catch problems before your customers do → You stop paying for agents that take 6 steps when 4 would do The whole system is open source. Free to use. If you're building AI agents for your business, evals is how you verify.

- Drafted a blog post - Used an LLM to meticulously improve the argument over 4 hours. - Wow, feeling great, it’s so convincing! - Fun idea let’s ask it to argue the opposite. - LLM demolishes the entire argument and convinces me that the opposite is in fact true. - lol The LLMs may elicit an opinion when asked but are extremely competent in arguing almost any direction. This is actually super useful as a tool for forming your own opinions, just make sure to ask different directions and be careful with the sycophancy.

@steipete This is crazy

This guy emailed me asking for a *token session refund* because his claw made mistakes. 🙃

@sweatystartup Until you host your own model and train it. I know it’s easier said than done but company will start doing it once they will be charged actual cost of ai model’s after ai model war is done.

Investing heavily in AI at your company will backfire. You are becoming dependent on something that is unsustainable. The VC money will dry up once they realize nobody is going to make any money in the long run except NVDA and the power companies. The subsidies will stop. And your costs will 5x. There is no moat in AI. Switching from GPT to gemini to grok to claude takes seconds and you don't miss a beat. Its a house of cards.

@phosphenq My men.

> you spend an 3 hour crafting the perfect tweet > rewrite it three times > post it at what you think is the right time > 34 likes. 2 retweets. zero quotes > check it an hour later hoping the algorithm picked it up > it didn't > meanwhile someone quote tweets an article with a 15 second movie clip > 8 words. 0 original content. 10 million views > they do this every day. different clip. same formula > the algorithm that decides who gets seen is open source > an AI agent can read your feed and detect the exact format pulling millions > you were never bad at content. you just didn't know the system

@JamesEbringer Hey James dont see a dm yet

@JamesEbringer Stunt

One person is managing 20 AI UGC pages simultaneously Posting 2x a day on each one Sounds impossible but each video literally takes 45 seconds to create I built the tool behind it Reply "STUNT" and I'll send you access to it

@Hartdrawss Easier said than done. Marketing is the hardest part than building.

The easiest path to a viral indie app right now (that actually makes money): > go to appkittie[.]com → filter apps making over $50k/mo > sort by rating (low to high) → pick the worst rated ones > open the app store page → scroll straight to 1 and 2 star reviews (this is the goldmine) > copy-paste the 20-30 worst complaints into Claude with this prompt: "summarize core user problems, group by theme, quote exact pain phrases, ignore bugs, focus on missing features / bad UX" > you now have a crystal-clear list of what users hate about an app that already makes real money > build the 80% better version that fixes the top 2-3 complaints (simpler UI, one killer missing feature, faster flow, no forced sign-up) > launch fast → market on Reddit, TikTok/Reels, X threads $800-1.5k MRR in 60-90 days → flip for $8k-15k on MicroAcquire find demand, then build

Repeat after me dont ever give articles directly to your openclaw or any ai agent. It will mess up your setup, will cost hours and model tokens and you will be ended up frustrating fixing it.

Me reviewing the code written by Claude before pushing it to production

Drowning in the AI slop. It seems like everyone is making killing with OpenClaw agents while I am just "using it" Every one is minting millions using polymarket agents and I am on the sideline watching them. Everyone has 1000s of followers while typing random tweets while I am thinking about every single tweet that I send. Feels frustrating.