Ajay krishna A

Keralam is truly leading by example after PM Shri @narendramodi ji called for reducing vehicle usage due to the Middle East situation. Earlier, our CM used to travel with a huge convoy, causing traffic blocks and fuel waste everywhere. Now there’s no convoy. Also, we don’t even have a CM.

@ankitatIIMA @Akshat_World

💥Only 4 years late. Translation - Only place where there will be maximum freedom for citizens will be Bharat. The West will begin entering Digital prison cities to curb their consumption because their consumption based on free and endless printing of currency is coming to an end.

Dear @BJP4Bengal now that you are in power, please crackdown the huge network of scammers operating from West Bengal. They are working hard to ruin India's honor day in and out.

@AshwiniVaishnaw ji, you told us that starting January 2026, we can change train ticket dates directly through the @IRCTCofficial website. But this feature is not yet available, atleast I cannot see. I know you're capable.. so please can you make this available by tomorrow/ASAP?

While Gujarat builds semiconductors and Bangalore does deep tech, this is the kind of innovation that happens in Kerala.

@UlaaBrowser Hi, came back here to congratulate you guys on listening to me and doing this. Thanks @UlaaBrowser.

@UlaaBrowser Hey @UlaaBrowser team, the browser is not designed for dark mode. the navigation/address bar and the tab/title bars are the same as light mode.. please redesign to for dark mode also asap. this should have been done way earlier.

Meet Arun Justus and Diana Loomis from the Ulaa team at DEF CON Chennai 📍 Get a closer look at features designed for secure, managed work environments, and discover how Ulaa Enterprise brings Zero Trust security and granular control to enterprise browsing.

Why does @antigravity show my recent conversation way down in the history section of antigravity editor? @Google?

@warikoo @karpathy @sriramk @warikoo you are using paid tier claude cowork. its still split between the have and have nots.

Yesterday: A 22yr old commented “I don’t use ChatGPT because it makes up facts and the I have to spend time checking it on my own” Also yesterday: I got Claude cowork to draft appraisal emails for my entire team, based on their performance across the year plus the budget I gave it, personalizing the feedback to a degree that felt eerily me! The world was earlier split between the have and have nots. Today it is split between the don’t want to and the want to. Crazy times!

Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code. But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along. So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions. TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.

This looks so plastic. Feels like it was made on Higgsfield AI.

@airtelnews But airtelblack.com.

Bharti Airtel, the world’s second largest telecom operator by customer base, has crossed 650 million customers globally.

@malhotra_namit Bro this was so bad! I got cheap serial vibes, Ranbir as Rama didn't make that punch and it was AI slop all over the place. You made it feel like Game of Thrones, instead of being authentic. Take your time brother and do better!

Rama is the greatest of all time because he lived a life where the choices he made were always in the benefit of the greater good, duty over desire, and sacrifice over self. His legacy continues to enhance and empower humanity over time and bring the belief in the goodness of the human spirit to resolve all conflict and bring peace to the world. Namit Malhotra’s Ramayana Directed by Nitesh Tiwari In cinemas worldwide Ramayana Part 1 Diwali 2026

@MeghUpdates Bro is a mallu?

Kerala has talent. That was never the problem. Problem was no enough opportunities here. So people left. Gulf, Bangalore, wherever work was there. Now imagine this: You start a business → you create jobs → government supports you. ₹1 lakh for every job you give. 100 jobs = ₹1 crore support. That’s not small. That can change how businesses grow in Kerala. #NDAKeralamJobsRevolution 6 major municipal corporations. Each with a clear focus, not random development. TVM → tech & innovation Kochi → ships & marine Kozhikode → healthcare Plus big industrial zones, global companies, better skill training. Not just talk… actual plan to keep Malayali youth here. Led by Narendra Modi, with people like Rajeev Chandrasekhar and Sabu M. Jacob who understand business. Simple question: Next generation Malayalis… will they build their life here, or still leave? That answer depends on what we choose now

@shashank_kr @Mastercard @Visa why not rupaycard??

We are announcing the launch of Passkey, a biometric authentication for card payments that’s fully compliant with RBI's two-factor authentication framework. We launched this industry-first partnership with @Mastercard and @Visa. Almost a third of payment failures on cards in India trace back to OTP friction; delayed delivery, wrong entries, redirect loops. This friction literally runs into hundreds of crores in lost revenue, abandoned carts, and broken trust at the exact moment a customer decides to buy. But now, there’s no waiting for an OTP. Your fingerprint or your face and authentication & transaction is done. The authentication happens on your device, card data stays tokenised, and the whole experience becomes invisible at checkout. @Razorpay’s biometric authentication delivers up to 95% transaction success rates. For businesses, that’s a great relief on the last mile.

@anurag28bansal Your content is amazing Anurag! I'm lovin' it.

A huge chapter in my life comes to a close today. After 3 years, I’ve decided to move on from Zerodha. This is not easy to write. The last 3 years have changed me, personally and professionally, in ways I genuinely could not have imagined for myself. A very, very large chunk of that credit goes to the kind of support Zerodha gave me. Now, what’s next for me? I’m not fully sure yet. What I do know is that I want to go all in on content. YouTube, Instagram, and whatever else makes sense along the way. I’ve been doing this alongside my job for a while now, and it felt like the right time to give it everything I have. Over time, I hope to build something meaningful in the media or marketing space. I don’t have a big plan yet. If you’ve been following my work so far, thank you. And if you’re just discovering it now, I hope you like what I do! More soon.

@deepigoyal "It's that our CFD studies, aerodynamic models, SIL simulations, and flight logs all match, almost perfectly." That's great! keep rocking

We just completed a successful flight for our tech demonstrator, Lat One v0.2. Lat Aerospace is building ultra-short takeoff and landing aircraft using blown wing technology. Think fixed-wing performance with near-helicopter access. A couple of months ago, v0.1 achieved uSTOL but crashed shortly after. That was expected. v0.2 was about completing the full mission, and it did. The blown wing concept worked in reality. Closed-loop control got validated. We'd predicted a cruise speed of 30-32 m/s and cruised comfortably at 33. No thermal issues despite a burning afternoon. The quick-detach wings held through aggressive turns. Was in the air for over 6 minutes. Smooth touchdown, and ready to fly again. But the thing that makes me the happiest isn't any of that. It's that our CFD studies, aerodynamic models, SIL simulations, and flight logs all match, almost perfectly. That's what real engineering looks like. I am so proud of our team. A long way to go, and we are getting there. @lataerospace @surobhidas

@ArtificialAnlys @SarvamAI Can you provide a benchmark on performance per Billion parameters?

India enters the open-weights AI race with its largest models pre-trained from scratch: Sarvam 105B and Sarvam 30B @SarvamAI's Sarvam 105B and Sarvam 30B score 18 and 12 on the Artificial Analysis Intelligence Index respectively. Announced at the India AI Impact Summit 2026 and open-sourced under Apache 2.0, both are Mixture-of-Experts models trained entirely in India using compute provided under the IndiaAI Mission (@OfficialINDIAai). Both support reasoning and non-reasoning modes. These are an improvement from Sarvam's previous model, Sarvam M (8 on Intelligence Index, 23.6B parameters), which was based on Mistral Small rather than pre-trained from scratch. Sarvam 105B has 106B total parameters with ~10B active per token and a 128K context window. Sarvam 30B has 32B total parameters with ~2.4B active per token and a 65K context window. Alongside the text models, Sarvam also announced Saaras v3 (Speech to Text) and Bulbul v3 (Text to Speech) with a focus on Indic languages. Key takeaways in reasoning mode: ➤ Sarvam 105B scores 18 on the Intelligence Index. Among ~100B-class open-weights reasoning models, it trails GLM-4.5-Air (23), INTELLECT-3 (22), Mistral Small 4 (27), and gpt-oss-120B (High, 33). All four peers also activate more parameters per token ➤ Sarvam 30B scores 12 on the Intelligence Index. Among ~30B-class open-weights reasoning models, it trails GLM-4.7-Flash (30), Nemotron Cascade 2 30B A3B (28), Qwen3 30B A3B 2507 (22), and Qwen3 32B (17). Sarvam 30B activates fewer parameters than these peers. ➤ Sarvam 105B's relative strength is in select agentic tasks. Its agentic index of 25 places it ahead of INTELLECT-3 (20) and GLM-4.5-Air (21) despite trailing both on overall intelligence. Its GDPval index of 773 also edges ahead of GLM-4.5-Air (665). Both new models are a large step up from Sarvam M (Reasoning), which scored 8 on the Intelligence Index. ➤ Compared to peers, both models score lower on TerminalBench Hard (Agentic Coding & Terminal Use) and AA-Omniscience. Sarvam 105B scored 1.5% and Sarvam 30B scored 2.3% on TerminalBench Hard, compared to GLM-4.5-Air (20.5%) and INTELLECT-3 (9.1%). The AA-Omniscience Index is -60 for Sarvam 105B and -72 for Sarvam 30B. Both models have high hallucination rates relative to their accuracy, and both attempt to answer far more questions rather than abstaining, which drives the negative scores. Key model details: ➤ Modality: Text input and output only. ➤ Context window: 128K tokens (Sarvam 105B) and 65K tokens (Sarvam 30B). ➤ Pricing: Currently free on Sarvam's first-party API. ➤ License: Apache 2.0. ➤ Availability: Sarvam's first-party API; weights available on @huggingface and AIKosh.

@DeryaTR_ 3 is??

My new Sunday morning routine: 1. Get coffee 2. Check GPT-5.4 projects on the Codex App, continue & start new ones 4. Launch ChatGPT 5.4 Pro for fresh brainstorming sessions 5. Think/learn how to use the 90% of AI capabilities I have yet to explore 6. Drink more coffee

@protosphinx @pratykumar .

Whenever I bring this up, the pushback is that it’s just about money. I don’t think it’s that simple. Money is clearly important but this goes deeper. This is art. No one carves veins into stone without real dedication to the craft. I hope we see tech design (not merely design but tech + design) from India with the same level of craft.

@DataGuyRedux @grok does today's datacentres also need high amounts of water?

This is the same amount of water, which will produce 38kgs of Sugar.