Aman Hanspal

Hmm…

@JUNCRIS36371756 @jwrld1808 Dhoom 2 on Netflix

@jwrld1808 What's d title of the movie?

Bro literally disguised himself as a statue to pull off a billion-dollar diamond heist. 🤯 Absolute masterclass.

Gemini Omni is a major leap in world understanding & multimodal editing! It can take photos, video & audio and build entirely new scenes. Over time it’ll be able to handle any input & any output - starting w/ video You can even give it your own videos & iterate on your ideas:

@sudhakarp1x @IndianTechGuide Local/On-device AI + banking could. Financial data is extremely personal and private- high stakes, big banks wouldn’t want to risk users data or their internal personal company related being processed over servers

@IndianTechGuide AI + banking can change the future, but privacy and security should be the top priority when linking bank accounts with AI

🚨 ChatGPT can now be linked to bank accounts, will track your spending and savings. (India Today)

@ramxcodes @TheArnabSaha @Athaxv05 @ayan_duttait @hiarun02 @ayush__64 @tiw_ari_ayu I vote for number 44 @debsourya005 👀

41. Arun Kumar / @hiarun02 Project: snippify.live Launch: x.com/i/status/20406… 42. Ayush Sharma / @ayush__64 Project: …growth-production-d4e5.up.railway.app Launch: x.com/ayush__64/stat… 43. Ayushman / @tiw_ari_ayu Project: vibefix.site Launch: x.com/tiw_ari_ayu/st… 44. Debsourya Datta / @debsourya005 Project: github.com/debsouryadatta… Launch: x.com/debsourya005/s… 45. Deep panchal / @deepp2108 Project: kanvas.website Launch: x.com/deepp2108/stat… 46. Dev Chauhan / @devchauhann3 Project: reicon.dev Launch: x.com/devchauhann3/s… 47. Dhruvil Mistry / @dkmistry0203 Project: github.com/dhruvil0203/nu… Launch: No launch video 48. Divy Sharma / @Divy_Sharma6 Project: github.com/divysharma7/fi… Launch: No launch video 49. kartik sharma / @devkartik17 Project: clarixai.in Launch: x.com/devkartik17/st… 50. Krish Gupta / @_krishgupta5 Project: openforge.in Launch: x.com/_krishgupta5/s… 51. Nikhil Rajpurohit / @nick_realm_01 Project: github.com/nikhil008-git/… Launch: x.com/nick_realm_01/… 52. Nishul Dhakar / @nishuldhakar Project: xarti.drova.studio Launch: x.com/i/status/20522… 53. Pranav Vijay Gawande / @pranvv27 Project: devjobs.site Launch: linkedin.com/posts/pranav-g… 54. Priyanshu Shekhar Singh / @priyanshuwq Project: Closed source Launch: x.com/i/status/20530… 55. Saad Haider / @_MSaaDH Project: A Redis-style in-memory server in Rust Launch: x.com/i/status/20553… 56. Saeed Saiyed / @saeedsaiyedtwt Project: img0.site Launch: x.com/saeedsaiyedtwt… 57. Sahil Mishra / @SahilMi86109458 Project: ghostmark.sahilmishra.dev Launch: x.com/SahilMi8610945… 58. saurabh garkoti / @_saurabh__xd Project: memehunt.tech Launch: x.com/_saurabh__xd/s… 59. SUKHJIT SINGH / @thesukhjitbajwa Project: beta.resumely.tech Launch: x.com/thesukhjitbajw… continue below 👇

#100_cans community voting is now live. > VOTING WINDOW From: 17 May 2026 — 8:00 PM IST To: 18 May 2026 — 8:00 PM IST > VOTING FORMAT I vote for Number Example: I vote for Number 21 Rules: > One vote per person. > If you comment multiple times, only your latest vote counts. > Votes can be: only under this post Just keep the format exact. (i’m scraping this automatically, so don’t try creative formats 🙏) Participants can now share their number and ask people to vote for them. (in QRT's) --- PROJECTS --- 1. Aditya Shukla / @AdityaShuk87206 Project: github.com/ShadowAdi/term… Launch: x.com/AdityaShuk8720… 2. Agastya Khati / @krish725_ Project: github.com/kris70lesgo/He… Launch: x.com/krish725_/stat… 3. Ashutosh Singh` / @ashudotdev Project: devmirror.ashu.fun Launch: x.com/ashudotdev/sta… 4. Bhavya Goyal / @Bhavyaztwt Project: rustvisual.netlify.app Launch: x.com/Bhavyaztwt/sta… 5. Govind Nagar / @Govindnagar__ Project: avtar.studio Launch: x.com/i/status/20553… 6. Ketan Kauntia / @KauntiaKetan Project: patchesplay.com Launch: x.com/KauntiaKetan/s… 7. Manish Kumar / @Manixh02 Project: github.com/ig-imanish/ope… Launch: x.com/i/status/20551… 8. Mitali Sinha / @kayleecodez Project: CLI Tool Launch: x.com/kayleecodez/st… 9. Nijeesh NJ / @nijeeshnj Project: github.com/codebyNJ/AdenV… Launch: x.com/nijeeshnj/stat… 10. Prince yadav / @prncexe Project: npmjs.com/package/rebar-… Launch: x.com/prncexe/status… 11. Punyakrit Singh Makhni / @punyakrit_22 Project: qrise.st Launch: x.com/punyakrit_22/s… 12. Sushant / @iamcodeblooded Project: x.com/iamcodeblooded… Launch: x.com/iamcodeblooded… 13. Ufraan / @ufraaaan Project: github.com/ufraaan/tex-le… Launch: ufraan.dev/projects/tex-l… 14. Vaibhav Gupta / @ignvaibhav812 Project: github.com/ignvaibhav/Fer… Launch: x.com/ignvaibhav812/… 15. Vinayak Ahuja / @vinayakahuja64 Project: papaai.netlify.app Launch: x.com/vinayakahuja64… 16. Abhay Parekh / @abhxy03 Project: flow-stack-studio.runable.site Launch: x.com/abhxy03/status… 17. Abhijit jha / @abhijitwt Project: x-algorithm-six.vercel.app Launch: x.com/i/status/20553… 18. Achintya K J / @ciatelarlanta Project: chroma-shift-ten.vercel.app Launch: x.com/ciatelarlanta/… 19. Aditya Chaudhary / @useradityaa Project: nexus-world.runable.site Launch: x.com/userAdityaa/st… 20. Aman Jaiswal / @Aman09143227 Project: gitlense.runable.site Launch: x.com/Aman09143227/s… Continue below 👇

I vote for number 44 @debsourya005 👀

@Prof_Cheems @Divanshu___Garg

So cute 🥰

@SBallaya back when TPC was gold 🥲

RAIT 🫪

want to get started with JEPA for world models? don’t start by reading 20 papers. start by understanding the core idea. what jepa actually is: • instead of predicting pixels, it predicts embeddings • the model learns a latent representation of the world • it tries to predict “what representation should exist next” • less focus on raw reconstruction, more focus on abstract understanding why this matters: • pixel prediction wastes compute on tiny details. • world models need structure, dynamics, and relationships; not every pixel value. how to approach it: • learn the prerequisites first → transformers, self-supervised learning, contrastive learning, representation learning • understand embeddings deeply → vectors as compressed meaning • study predictive learning → context encoder predicts target representation • play with video datasets → motion and temporal consistency matter • implement tiny versions → don’t start with giant models. train on simple sequences first • focus on latent space behavior → visualize embeddings, transitions, clustering what to study around it: • Self-Supervised Learning • Contrastive Learning • Transformer • World Model jepa is part of a bigger shift in ai. models are moving from “predict the next token/pixel” to “understand the structure and dynamics of reality.”

imo... GPT-5.5 feels more sugarcoated — agreeing with whatever you say. Opus 4.7 feels like a straight slap to the face: grounding you whenever and wherever needed. (this is general discussion and brainstorming, not coding)

Real-time World Models are the next AI frontier. Today, we @reactorworld are taking the first step towards this reality: our early preview lets you experience worlds generated in real-time, running on our global low-latency infrastructure. Try it now: reactor.inc

Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens - Less than 5% the cost of Opus Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention). Only a small fraction actually matter. @subquadratic finds and focuses only on the ones that do. That's nearly 1,000x less compute and a new way for LLMs to scale.

@VazeKshitij @raghav_chadha

@raghav_chadha Nah man😂😂😂😂

Congratulations to the entire BJP family on the impressive election results, with West Bengal marking a historic & decisive victory that stands out as truly remarkable. The outcomes from Assam and Puducherry further reflect people’s continued faith in development and strong governance. This victory stands as a testament to the decisive leadership of Hon’ble PM Shri @narendramodi Ji, the strategic vision of Hon’ble HM Shri @AmitShah Ji, and the dedicated efforts of BJP National President Shri @NitinNabin Ji. A win for every citizen aspiring for a stronger, more developed India 🇮🇳

@AvinashSingh_20 Mumbai?

Bangalore Job seekers Kit Sheet link- #gid=116538532" target="_blank" rel="nofollow noopener">docs.google.com/spreadsheets/d… Startup Bundle- topmate.io/letscode/20711… .

Normalize disappearing for an extended period of time to do a factory reset on your focus and creativity.

Some observations on agents from a lot of time spent installing, using, debugging and building things with OpenClaw, Hermes, etc. - Most people who rave about OpenClaw, HermesAgent etc. have not actually used it. Facts. - Agents are developer products today. They are so so so far from being consumer grade in terms of delight, simplicity, reliability etc. - I wish the agents generated more dynamic interfaces. Don't give me chat. Give me interactive, dynamic apps/webpages. I suspect this will be a next big vector of innovation. - The agents are ultimately a wrapper to code-generation and tool-calling....not to "generative AI". This is important to grok. These agents are doing most of their "work" through iterative tool calling. - Once you understand the core loop behind agentic AI, old school "chat AI" doesn't make sense. The idea of an "always-on" agent that actually adapts, learns and maintains state is a LOT more compelling than a one-off chat prompt. The latter feels much more like a search query. - So much of mangling with agents today is basically setting it up to use browsers and/or your computer. And there are so many constraints! - Installing and running openclaw feels a lot like running Linux in the early 2000s. You spend a lot of time compiling device drivers, setting up configurations etc. You feel a real sense of accomplishment from jerry-rigging everything together but it is a LOT of work to get to the point of utility

@ryandeiss @grok Aight, thanks!

@han__ama @grok You can also look up his talk about innovator’s dilemma.

Smart move by Apple. People will call this "succession planning," but what's really happening is way more interesting… This is a corporate rejuvenation cycle and it's a pattern that plays out at every level of business. Every company goes through three leadership phases. Phase A: You need a Visionary to build the thing. Phase B: You need an Operator to scale the thing before it crashes under its own weight. Phase C: Once the thing gets "boring," you need a Builder (Visionary 2.0) to make it exciting again before a competitor disrupts you. Apple just moved from Phase B to Phase C. Steve Jobs was the visionary because he invented the thing. Tim Cook was the operator because his background was: - Industrial engineering - Supply chain - Operational efficiency He took what Jobs built and scaled it from $350B to $4T in market cap. > Revenue nearly quadrupled > Installed base grew to 2.5 billion devices > Services became a $100B business on its own That's what top-notch operators do. But at a certain point, simply making the product "better" no longer moves the needle the way better operating and selling does. The product starts to feel safe, predictable, optimized, etc. The company loses its product pulse, its understanding of the customer, and eventually everything stagnates. (Steve Jobs actually talked about this. Google it.) By picking John Ternus, a hands-on engineer who spent 25 years building the actual products, Apple is signaling that the optimization era is over and it’s time to build whatever comes next. This isn't new or revolutionary. Microsoft did the same thing. Bill Gates was the visionary builder. Steve Ballmer was the operator. Focused on sales, execution, and profit. But during that era, the product DNA was sidelined and Microsoft became a punchline. Then Satya Nadella came in… …an insider who understood the "how" of technology, not just the "how much" of profit. He moved the focus back to products like Azure and cloud and rebuilt the culture of the entire company. Microsoft went from punchline to $3T. Visionary -> Operator -> Builder -> Operator -> Builder (repeat) This same cycle plays out at $5M -> At $15M -> At $50M. The founder who's great at building the product but terrible at operations needs their "Tim Cook phase." They need systems, discipline, and someone who can scale what they built without it collapsing. The founder who's been in operator mode so long they've lost the creative spark that made the business special in the first place needs their "Ternus phase." They need to bring the builder energy back before the market moves on without them. Business is a constant balance between invention and operation. Apple is just demonstrating it at the largest scale in history. So, which phase is your company in right now? And do you have the right leader for it? (If not, are you willing to become the right leader for it?)