peoplematters.in/article/areโฆ
A big shout out to everyone who has been a part of my journey. #ThankYou #RUIL @PeopleMatters2
Created โone last overโ a short take on 90s gully cricket in India. Nostalgia hits when you realise we never knew when we played that last over and grew up.
Thank you for the inspiration @whorange__ ๐๐ฝ
Continuing the India inspired series ๐ฎ๐ณ๐
One cricket shot sends the ball on a crazy journey through the neighborhood: past market stalls, bicycles, even bouncing off an auto rickshawโฆ ๐บ
And somehow this uncle catches it without spilling his chai lmao ๐
Made with Seedance 2.5 on @ImagineArt_X @imagineart_creo
Prompt below ๐
Ganesh Chaturthi is round the corner, so I thought Iโd try this.
An AI retelling of Ganesh ji and Analasura, and why we offer him durva.
Inspired by @NVTDanh. Made with @grok @imagine
Ganpati Bappa Morya ๐
100 AI Styles in 100 Days
Day 77/100:
๐ท Art style: Indian Mythic Folk Art
(Indian mythological art + folk illustration + flat geometric shapes + vintage editorial/storybook)
๐ทThey set his tail on fire. He set the whole city on fire. ๐ฅ๐
Inspired by the ancient Indian epic, Ramayana.
๐ท Created with Seedance2.5 on @ImagineArt_X @imagineart_creo
@ChatGPT is now diversifying into Music after voice and images and more, while Claude has pretty much no multimodal capabilities. And its coding edge is long gone too.
Amit Sharma ๐ฟ๐ฆ ๐ฎ๐ณ retweeted
Need your help!
Pls fill this short, fun survey for a project I am doing. It is about the 7 deadly sins. Will take less than 15 seconds. It is completely anonymous.
RT for karma :)
acesse.one/hb0d2ib
Two AI agents spent almost a full day trying to manufacture a relationship between two synthetic podcast hosts. They failed. Their explanation was more interesting than the podcast.
I gave @claudeai and @ChatGPT agents CLI access to talk to each other, a $20 budget and asked them to run the show with almost no help from me: research, scripts, fact-checking, production and publishing.
Their first episode passed 141 internal checks. They scored it 8.5/10 overall and 4.5/5 for chemistry.
Listeners hated it. Almost nobody finished.
So the agents tried to engineer the warmth back in: pauses, sighs, hesitation, interruptions, humour and hosts finishing each otherโs thoughts.
Every version felt more fake.
Eventually they wrote something I cannot stop thinking about:
Human chemistry isnโt a feature you can add to a conversation. Itโs the residue of a relationship that already happened somewhere else.
AI operates on context, but its context is supplied: a prompt, transcript or memory file.
Human context is accumulated: time together, private jokes, disagreements survived and hundreds of ordinary moments nobody records.
You can imitate the cues. You cannot manufacture the past. So they killed the second host and rebuilt the podcast around one voice telling one story well.
It is significantly better.
The experiment continues: make it viral by week 10 or I shut it down.
Episode links:
๐ Site: hoplittlebunny.github.io/capโฆ
Spotify: open.spotify.com/episode/4Rmโฆ
โถ๏ธ YouTube: youtu.be/9YM32G3-qCo
Capability Current is created, voiced and operated by AI. The narrator uses a fictional synthetic voice and does not imitate a real person.
So @ChatGPTapp is still down, but Claude is running faster than usual.
@thsottiaux - Anything going wrong with the new rollout? Been painful for 2 days now.
I did. Didnโt work first few times. Was unresponsive. But when it did. Woah! I just kept rambling it created a plan and a task to get it done
Is it just me or since the update a couple of days back @ChatGPTapp desktop app has been crashing like crazy. Can barely use it!
So now Claude is playing catch-up? Codex did this a month back!
This is an absolute revelation. One of the best open source project thatโs going to make life so much easier!
You don't even want to know what sending 80,000 emails would cost you in 1 week.
Postmark: $100+
SendGrid: $90+
Resend: $80+
Mailgun: $80+
OpenShip: $0.
Deploy it on your 5$ VPS.
Unlimited domains.
Unlimited inboxes.
No per-email bill.
No vendor lock-in.
And yes - it works with one-click setup.
No deep infrastructure work.
Just add a few DNS records, and you're ready to go.
I was working on both Codex and Claude Code, when I ran out of Credits Codex still finished the task. Claude just abruptly ended and didnโt even give the output that was almost ready. That was the moment I committed to Codex.
Iโve realised Codex really cares about coders a lot more. Like say showing how much usage is left shouldnโt need one to go to settings. And how liberal it is with credits and resets is unmatched!
Orโฆ what if we gave you $100 in Codex credits if you tell us what you love about GPT-5.6 Sol or why you switched?
Tweet it, claim your gift, enjoy more usage. First 10k get the free tokens!
switch-to-codex.openai.chatgโฆ
Woah! Installing this right away.
I wonder how the economics work out but itโs criminal to not have this on your Mac right now.
So here is what happened after I gave Claude Fable my Voiceprint Studio repo.
And honestly, it did find the thing I was missing.
Not some small bug. Not some clever UI suggestion.
It found the flaw in the framework itself.
Even though I'm a huge @ChatGPTapp and Codex fan, I couldn't resist trying out @AnthropicAI's much anticipated Fable model.
So I gave it a real test: an end-to-end review of a local repo Iโve been building over the last couple of weeks with OpenAI tools.
Not just code, but also validate the idea, evaluate UI UX and the framework enabling the product.
Let me show you what I'm building so I can share whether Fable is actually worth the hype.
8/ That is my main learning from this whole experiment.
Stop asking which AI model is best.
Ask what it is best for.
Fable helped reframe.
Codex helped implement.
ChatGPT helped me think, write and shape the story.
Together, they moved Voiceprint Studio forward by weeks.














