Alex Jax

@byungkyu_p @HubSpot @Shopify @google @SlackHQ @Jotform @awscloud @NotionHQ Hey, nice app but how are you making sure our keys aren't leaked like it happened with @composio the other day?

We built a toolkit that connects your AI to any app in 3 lines of code. Comes with hundreds of pre-built actions for > @HubSpot > @Shopify > @google > @SlackHQ > @Jotform > @awscloud > @NotionHQ > + more. Works with MCP, OpenAI, AI SDK, and LangChain and takes care of auth (OAuth, API Key) See how it works ⬇️

@gurungbuilds @composio same here...

@composio saw an email revoking the github access and just read the such a let down, since i genuinely thought composio would be pretty safe w auth tokens and even suggested it to everyone i know

We've identified a security incident that involved unauthorized access to certain internal Composio systems, impacting a limited number of customers. We will share more as we learn more. Please see our security bulletin: composio.dev/blog/composio-…

@HermesAgentTips @NousResearch Oui :)

@alexjax @NousResearch Haha fr

It’s pronounced “Hermes”, not “Hermes”

@richschefren Rich, I switched... see my last post :)

@alexjax Very interesting. Is there really nothing better about Hermes?

Is Hermes really as good as people say? How does it compare to Open Claw. #ai #business

I switched from @openclaw to Hermes by @NousResearch last weekend after using OpenClaw daily since February, and I wondered why I waited so long. Using the same tool every day can normalize small problems, much like walking with a tiny pebble in your shoe. Longer coding tasks and jobs that required significant compute time were often delayed on OpenClaw, but that’s not the case with Hermes. The result? Fewer restarts, less babysitting, and more time for thinking. What changed in just 48 hours on Hermes: - Fewer distractions pulling my attention. - Large compute tasks completed without hanging. - My day now includes larger, quieter blocks for real focus. - I also built a three-layer memory system that enhances Hermes's capabilities. Here's how the memory system works: - Built-in Hermes memory: Lean and always injected, it stores stable facts like who you are and how you want me to behave. - Hindsight as active memory: Runs locally on my VPS, managed by systemd as hindsight-hermes.service, with a PostgreSQL store for automatic recall and retention. - QMemory / Qdrant archive: A curated long-term vault for project notes, research, and PDFs, with vectors in max_long_term_memory (29422 points of memory - that's a lot 🤣 ). Why this matters: Little annoyances can accumulate. Faster, more reliable compute, combined with a sensible memory stack, restores your attention and momentum. If you're considering a switch, try this simple test: - Use the new tool on real work, not a demo. - Count interruptions and manual fixes for 2–3 days. - If you gain more uninterrupted time than you lose in learning, stick with it. What surprised me most was the immediate return of time, without a slow relearning period. No nostalgia, just forward motion. What tool did you switch to this year that improved your day? Share one thing it fixed, and I will reply. P.S. If you want DM me to discuss your own personal / business AI Assistant.

True

@bcherny Good luck Boris. I see the room is vibing 😁

Hello from Code with Claude London!

@robby_mtf @papercliping That's exactly my plan with it 😀 a full marketing pipeline. Did you give the video agent skills like remotion?

@alexjax For sure! Think of @papercliping as an org rather than an employee - for me it’s a marketing org (cmo then under it agents for video, blogs, trends, etc). Many ways to shape it - some run entire businesses on it so the org would be headed by a CEO w/ functional agents underneath

Paperclip is going to replace all of my OpenClaw usage If you haven't tried it yet definitely check them out!

@dotta Still on the fence to moving over from @openclaw

@alexjax It's getting better every day

📎 Paperclip v2026.512.1 is out. Featuring 👼 Stronger liveness - your clips will work harder, plan better, and run longer 🧩 Plugin managed agents, routines, and skills 🔐 Secrets - w/ first-class provider vaults 🧠 Planning mode 🔍 New Search github.com/paperclipai/pa…

Before: manually copying customer details between 5 different systems and chasing updates. After: one click and everything syncs automatically. Automation isn't just for big companies. What's one task you wish took less time? #UKBusiness

@TommasoGritti @steipete Me too, if you find how, please share

@steipete 👀 I am all eyes/ears/prompts

Birdclaw has my complete twitter archive, so I can ask Codex for any old weird tweet I ever favorited or bookmarked. birdclaw.sh

Turn these meeting notes into a professional client proposal. Structure: - Project Overview - Scope of Work - Timeline & Deliverables - Investment (time or resources) - Next Steps Make it persuasive but professional, clear and benefit-focused. Here are the notes: [notes here]

Quick AI tip for UK small business owners: Turn messy meeting notes into a polished, professional client proposal in under 5 minutes. The right prompt makes all the difference. 👇 #SmallBusinessUK

@WhitoUK Absolutely true, thanks for adding that

@alexjax The time saving is real, but the bigger win is consistency. Templates and automation mean every client gets the same quality of communication, not just the ones you chase when you have spare capacity. That's what actually protects cash flow.

Before AI: spending 4 hours a week writing the same client emails and chasing responses. After: 15 minutes. The difference is ridiculous. Anyone else drowning in repetitive admin? #UKBusiness

@nikos1 Did you check grok argumentative mode? 🤣

Make it argue with me when I’m wrong, tired of models that just agree

@richschefren They are both the same. If you want a nicer interface then yeah, Hermes has one but functionality-wise they can both be set to do exactly the same thing. Ultimately you have to test both and choose what suits you.