few simple reasons why b2c is absolutely brutal:
- your cac is actually adversarial. you’re bidding against everyone else for the same eyeballs, & the platforms (meta, google, apple) all tax you (e.g. most startups are running creator ads or real ads). the platforms also periodically change the rules or the algo. b2b can ride sales relationships whereas b2c mostly *rents* attention.
- retention is completely unforgiving. consumers are fickle, switching costs are near zero. you’re sorta reearning the user every session which sucks.
- willingness to pay is low & emotional. enterprise will drop $200k on boring software but consumers often churn over a $2 price hike or a vibe shift.
- the distribution layer is all power law & mostly owned by someone else. discovery is winner take most & gatekept.
Sovereign AI at population scale isn’t theory anymore, it’s shipping.
Sarvam AI is building a full-stack, “Made in India” AI platform that:
🧠 Trains 100B+ parameter MoE models efficiently across 4,096+ NVIDIA H100 GPUs
⚡ Delivers millisecond-level, multilingual voice inference for Aadhaar
📞 Powers automated KYC, sales, and support across telephony and WhatsApp for brands like Tata Capital and Infosys
By integrating ASR, LLMs, and TTS into low-latency voice agentic workflows, Sarvam is bringing natural, real-time AI interactions to 1.4B Indians in their own languages. 🇮🇳
This feels like a blueprint for how countries can build sovereign AI without compromising on performance or safety.
Meet R2D2!
@tfadell recently tweeted that “the goal should be to use AI to accelerate learning, remove friction, and give people better tools to build expertise faster.” With this goal in mind, I built R2D2 using Codex.
R2D2 is a lightweight reading assistant built to simplify your learning experience! It resides inside your mouse cursor and acts as your learning companion.
Instead of copying text, switching tabs, opening a chatbot, and pasting copied text into the chatbot, R2D2 runs quietly in the background and appears only when you need it.
Here’s how it works: highlight text anywhere on your computer, in a browser, PDF, document, email, or an application, and R2D2 pops up to help you!
The selected text is already captured, so you can immediately ask a question like, “Explain this simply” or “What does this mean?” or “Give me useful context.”
R2D2 responds in real time and keeps the response clean and readable so that you can focus entirely on learning those tricky concepts that you are trying to master. It’s designed for fast reading, quick clarification, and staying in flow without leaving what you’re working on.
R2D2 is open source and available on Windows. Link to download R2D2 is in replies.
I was inspired by @FarzaTV's insanely great tool Clicky. While Clicky is more general purpose, I wanted something specific - in this case something for reading and learning - that I could run on my Windows computer. Clicky is an inspiring piece of work and I recommend checking it out (link in replies).
Come and visit London’s Home of Trophies. 🏆
Book your Stadium Tour at Stamford Bridge now. ⭐️⭐Come and visit London’s Home of Trophies. 🏆
Book your Stadium Tour at Stamford Bridge now. ⭐️⭐Come and visit London’s Home of Trophies. 🏆
Book your Stadium Tour at Stamford Bridge now. ⭐️⭐Come and visit London’s Home of Trophies. 🏆
Book your Stadium Tour at Stamford Bridge now. ⭐️⭐Come and visit London’s Home of Trophies. 🏆
Book your Stadium Tour at Stamford Bridge now. ⭐️⭐Come and visit London’s Home of Trophies. 🏆
Book your Stadium Tour at Stamford Bridge now. ⭐️⭐️
Andy Jassy on non-linear nature of AWS' journey.
"Progress jumps around; it’ll zig up, then
sometimes stall, or zag down, or force you back to the starting line. Sometimes, it feels like you’re running in
circles. But, the path is rarely straight."
#Amazon#AWS
@thekitze@ericzakariasson I remember reading a tweet that said that all the sub agents created by cursor are composer 2 agents (even if the model selected by you is gpt5.4 or opus 4.6)
Shouldn’t imagination rest with us humans? Fair to use AI for tasks and even brainstorming in certain cases. But surely we can’t offload our power of imagination to AI. Just my 2 cents. If anything, we should try to remove distractions from the reading process. This adds distractions.
@zarazhangrui I read this almost 12 years ago and there was a time when twitter was filled with such long forms. Sad that it’s hard to find quality long form writings.
https://t.co/TWy970v87e
Thank you @jacalulu for an excellent session on AI x Product Management during your office hours earlier this week.
I will for sure rely on your "4 directions" framework from here on while assessing new ideas.
Introducing the "Follow builders" skill: the best way to stay on top of the insane happenings in AI
I carefully curated 25 X accounts & podcasts that share the highest-quality, first-hand insights on AI (by builders from OpenAI, Anthropic, Google, OpenClaw, Replit, Vercel, Cursor...)
Your OpenClaw/agent can remix my central feed & send you a personalized daily newsletter in whatever channel you like
Already widely used with 2k+ stars on GitHub; link below