Priyanka

@vivoplt @karpathy @steipete @gregisenberg @rileybrown @jackfriks @levelsio @marclou @EXM7777 @eptwts @techhalla — For creative experiments, prompting, and generous sharing of resources @kimmonismus — AI news roundup. @NorthstarBrain — practical AI applications.

best 15 accounts to follow in AI: @karpathy = LLMs king @steipete = built openclaw @gregisenberg = startup ideas king @rileybrown = vibecode king @jackfriks = solo apps king @levelsio = startups king @marclou = startups king @EXM7777 = AI ops + systems king @eptwts = AI money twitter king @godofprompt = prompt king @vasuman = AI agents king @AmirMushich= AI ads king @0xROAS = AI UGCs king @egeberkina = AI images king @MengTo= AI landing pages king follow them all and learn.

@eng_khairallah1 Single-agent mode is like giving one person every job in a company. Multi-agent systems (with clear roles, handoffs, and shared memory) are where the real leverage happens especially for complex coding, research, or product work…

Boris Cherny, the creator of Claude Code at Anthropic, just explained why single-agent workflows are already dead in this talk he breaks down exactly how the future is teams of agents, not better prompts: - the 14% you lose to CLAUDE.md before typing a word - one agent researching. one building. one reviewing. one orchestrating - the architecture that separates hobbyists from real builders - the 3 properties every agent team needs to actually survive if you've been using Claude for more than a month and never left the chat window, you've been using one agent when you could be running a team of them instead of another show tonight, watch this make sure to bookmark it before it gets lost in your feed the guide is in the article below

@RoundtableSpace …Is this peak? Or is this the warm-up where in six months we all have digital clones that argue with our bosses better than we do while we just vibe in bed eating cereal straight from the box?!?

HERMES AGENT IS ONE OF THE FIRST AI PROJECTS THAT ACTUALLY REMEMBERS EVERYTHING ACROSS SESSIONS AND GETS BETTER THE MORE YOU USE IT. Multi-layer memory, self-evolving skills, autonomous 24/7 agents and cross-session recall. It feels less like a tool and more like an operator.

Terrified workers with no net? Blame slow policy, not builders. AI’s amplifying humans into superworkers (50%+ jobs reshaped, not erased per BCG). Productivity boom is inevitably incoming — cheaper everything, new demand verticals will be created. But yes the digital divide is going to put some countries far far ahead of others

Are billionaires actually insane? They’re dumping billions into AI, building data centers everywhere, laying off thousands of workers, and acting like none of this will have consequences. Everybody with a normal job can feel what’s coming. People are terrified. And there’s basically no safety net if it all collapses. It honestly feels like they’re racing toward some massive crisis while hoping we stay distracted until it’s too late.

@LazadaMY You guys know that you have the worst App UX right? The moment one gets in there’s a deluge of unwanted animated coupons that are so so so annoying …. The app needs some clarity of what and how you showcase your products. Learn from Amazon Prime

My locker is now 50% essentials & 50% things I bought during the May Payday Sale that immediately became essentials… 😅

Google Omni — 👏🏼🤌🏻🤌🏻👏🏼👏🏼👏🏼🤌🏻🤌🏻🎠🤓

@beastoftraal This is just a bad creative selling a goose 🪿

Adequately surprised to see Infosys using a plush-toy'ish cutesy mascot called Leon to sell AI to CEOs!! A mini-Yeti-like mascot that wears a "I heart CEOs" badge and is trying to directly persuading Fortune 500 CEOs on multi-million-dollar AI transformation decisions? That's mighty uncommon. 1/5

I wake up every morning to better ideas than I went to bed with. Not because I’m smarter. 😅Because I built this slightly ridiculous system I call Auntie Social. 👵🏻 I wanted to stop losing good ideas. My old workflow was hopeless: hear something smart → save it → never see it again. So I started dumping everything into one place. 👉🏻Voice notes from whenever. 🎠Screenshots. 🎙️Podcasts. 🎛️Half-baked thoughts. ⛓️‍💥X threads. 🔫IG memes. 🪎WhatsApp forwards. 🪤Weird meeting observations. ⛓️Links I swear I’ll revisit later. Every night the whole messy pile gets pushed through ChatGPT, Claude, and Gemini. I use all three because they’re weirdly different. Claude somehow pulls out emotional patterns and connections I missed. ChatGPT turns chaos into something usable. Gemini occasionally throws in a completely different angle that annoyingly turns out to be good. Then everything lands back into my Notion like three slightly opinionated aunties who stayed up gossiping about my life. I kept tweaking things with prompts, no-code tools, YouTube rabbit holes, and late-night experimentation until the workflow finally clicked. By morning I usually wake up to: 🪪old notes resurfacing at the perfect time, 🔩unexpected pattern matches, 🩻rough content directions, or sometimes one genuinely strong strategic idea.🧨 Last week it connected a podcast on creator loyalty, an old note about trader psychology, and a random affiliate conversation… then handed me a campaign direction I’m actually excited to build. Putting together a system that remembers and compounds better than I do has been a bittersweet journey! I’m increasingly convinced most of us are not short of ideas. We’re short of systems that let ideas collide again later. That’s probably why so much online content feels recycled now. Everyone consumes. Very few people reconnect. Anyway this is what Auntie Social looks like while I should probably be asleep linkedin.com/posts/priyanka…

Beyond the $42B trading shift, wondering abt secondary effects on the domestic ecosystem: • Thinner order books on Indian exchanges = wider spreads/premiums for retail users. • Reduced on-ramp liquidity means harder effort for genuine innovators (DeFi builders, tokenized assets, blockchain for remittances/supply chain). • Talent and startups drift offshore or to grey channels. … this high friction encourages “just hold USDT” behavior over productive use cases, turning crypto into a store-of-value escape hatch rather than a growth tool

India taxes crypto at 30% on gains and 1% TDS on every single trade But Binance told there are no requirements in any law… specifying withdrawal limits on virtual digital assets, They allow free withdrawals to private wallets. But they did not say it inside Parliament hall, Binance was invited to the Parliamentary Standing Committee on Finance meeting on 20 May 2026. Traditional remittances go through LRS with a $250K annual cap and 20% TCS above 10 lakh rupees. Crypto has none of these gates. Over $42 billion in trading volume shifted offshore since 2022 because the tax made staying more expensive than leaving, Toll booth on an empty highway. India has 100m+ crypto users and no crypto law. You can buy bitcoin on a registered exchange, move it to your own wallet, convert it to USDT, and sit on dollar value with zero LRS paperwork. Meanwhile if you want to send $1000 to your kid studying abroad, you fill forms, pay 20% TCS above 7 lakh, and wait for the bank to approve it. One channel is regulated to death. The other has no gate at all. The deeper problem is not tax evasion. It is dollarization through the back door. Every Indian who converts rupees to USDT and holds it is choosing the dollar over the rupee without the RBI knowing or being able to stop it. The DRI already found gold smuggling rings using USDT to move money to China, Global stablecoin supply crossed $316 billion in April 2026. India has no way to measure how much of that sits in Indian wallets, Cannot defend a currency you cannot track.

If crypto is a sneaky parallel system draining forex, why not integrate it with clear rules instead of half-measures? Or are we seriously trying to enforce LRS parity on decentralized ledgers… without turning off the internet? 😂 Regulation by confusion is the real stablecoin. #CryptoIndia

PM forgot to mention this in his speech. Buy crypto. No TDS, No LRS. Sell abroad or hold in USD stablecoin. This will drain forex big time even while traditional remittances go through hoops. Good story by @sugataghoshET

Yeah, it’s wild that one guy is about to pull in almost a billion dollars a year in dividends alone. But the bigger takeaway for investors: the company is mature enough and confident enough in its future cash flows to do this. Most hyper-growth tech names (think early Amazon or Tesla) never touch dividends because they reinvest every penny. NVIDIA just said “we can do both!!!!”

THIS IS INSANE. Jensen Huang will now earn nearly $870 million every year just from $NVDA dividends. NVIDIA just raised its quarterly dividend from $0.01 to $0.25, a massive 25x increase. Because Huang owns 871.7 million NVIDIA shares, his annual dividend income jumps from $34.8 million to almost $870 million in pure cash flow.

@AIandDesign @Meta Meta's not alone in this (we've seen it at Google, Amazon, etc.), but damn. The tech is exciting but watching big AI companies treat their own workforce like disposable training fodder is bleak

I'm not gonna lie, the @Meta layoffs are some of the most dystopian I've ever seen. They got told to work from home, they were sent the emails at 4AM in the morning. Those who weren't impacted have software on their computer that tracks their every move, preparing AI to take their job as well. They're literally training the AI that will eliminate their position as well. Meanwhile, Meta is raking in RECORD PROFITS. I am a massive, unapologetic AI enthusiast. Yet, this is NOT the future I had in mind. I wish for Meta to crash and burn. This is not the way. Literally nobody benefits from this.

This is how I adapted for myself -- 1. Used Claude Code (even Cursor, LangGraph will do) with separate .md agent files for cleaner handoffs. 2. Add validation gates (e.g., "Only pass if facts are sourced") and compaction between steps. 3. Start simple: Test on one content type you already know well (for me it was long form). 4. Iterate with feedback loops (e.g., add a Critic agent).

Anthropic's Head of Product just dropped a 28-minute masterclass on agent production. Prompt caching. Tool search. Programmatic tool calling. Compaction. Advisor strategy. 28 minutes. Free. Worth more than 100 YouTube videos combined. Watch it first. Then read this. The masterclass teaches you how agents work. This teaches you what to build with them — a 5-agent content pipeline that does the work of a $300K creative team. Full pipeline below ↓ Bookmark this. Start this weekend.

In your 20s, Urdu poetry feels romantic. In your 40s, you realise Ahmad Faraz was just describing adult scheduling conflicts. “Ab ke hum bichhde toh shayad kabhi khwabon mein milein…” Translation: “If we cancel this dinner again, we are not meeting until next financial quarter.” @priyanka.joshi_64413/nobody-warns-you-that-adulthood-feels-like-urdu-poetry-a2199fd4c374" target="_blank" rel="nofollow noopener">medium.com/@priyanka.josh…

The tech for faster strategy building/analysis is legit. All this “set it up in 10 mins, print $5k–8k/day” story is classic hype to drive Telegram joins. Trading bots (AI or not) lose money for most people due to overfitting, changing markets, fees, and psychology. If it sounds like easy money, treat it as entertainment…not financial advice

I like your confidence in humanity!!! 1- Resources like this have never been more accessible, which does lower the barrier to entry. 2 - You can go from zero to training small models pretty fast with free tools. But turning that into elite-level expertise still demands talent, obsession, and real-world mileage. If you’re actually starting - then do this 👉🏻the lecture + Andrej Karpathy’s “Let’s build GPT” videos + Sebastian Raschka’s book/code are excellent next steps. Just don’t expect a job offer after one video. 😄

Anthropic pays $750,000+ a year for engineers who know how to build LLMs from scratch. Stanford just released the exact lecture that teaches it - 1 hour 44 minutes, free, straight from CS229. Bookmark this & give it 2 hours today. It'll teach you more about how ChatGPT & Claude actually work than most people at top AI companies learn in their entire careers.

@Nithin0dha It's low-hanging barbed wire. Fix the basics (seamless banking, no courier KYC, simple taxes) or stop pretending diaspora money is welcome. Patriotism doesn't survive endless kyc + FX risks.

One thing that feels under-discussed in all the conversations about attracting foreign capital into India is the Indian diaspora. There is a large population of NRIs who are emotionally and financially interested in investing in India. But today, for many of them, the process of opening accounts, completing documentation, and actually investing in Indian markets is still far more painful than it needs to be. Making life easier for NRIs could be one of the lowest-hanging fruits for attracting long-term capital into India. This is something we’ve been focusing on heavily at @zerodha as well. Over the last year or so, we’ve made several changes to make investing as seamless as possible for NRIs. But there are still many frictions that exist because of regulatory and compliance requirements. Hopefully, SEBI and the government look at this more closely and think about how to make it easier for NRIs to bring money into India and participate in Indian markets. For a country trying to attract global capital, the Indian diaspora seems like the most obvious place to start.

I see this now as a narrowing gap between "I know what the software should do" and "I can make the software do it" but only when the person already carries some mental scaffolding from prior exposure. I would say real test will be whether this tool stays useful in 3/6 months from now

THIS IS CRAZY. Non-technical founders are officially dangerous now. A non-developer spent 30 days building with AI: 200+ hours. 70,000 lines of code. 30% tests. 15 people using it internally. 4-6 hours saved per person every week. But the crazy part is not the code. The crazy part is the system: → PRD before touching code → one micro-feature at a time → fresh session for every task → Claude for UI/UX → Codex for complex refactors → AI-written tests → Git for every rollback → security reviews before merging → never let 2 agents touch the same feature Vibe coding is not “prompt and pray” anymore. It’s becoming a real operating system for building software.