Subhankar

@OpenAI OpenAI straight-up acquiring Tomoro to drop 150 deployment engineers into the new company on day one? That’s a boss move. They’re not just shipping models anymore - they’re making sure big companies can actually use them.

We’ve also agreed to acquire Tomoro, which will bring 150 experienced Forward Deployed Engineers and Deployment Specialists to the OpenAI Deployment Company from day one.

Today we’re launching the OpenAI Deployment Company to help businesses build and deploy AI. It's majority-owned and controlled by OpenAI. It brings together 19 leading investment firms, consultancies, and system integrators to help organizations deploy frontier AI to production for business impact. openai.com/index/openai-l…

@cerebras @cognition Building multi-agent systems daily and Claude’s speed kills my vibe every time. SWE-1.6 looks insanely fast and actually agentic. Would use the $250 to go hard on a new open-source project that needs rapid iteration. Count me in!

🎁 We're giving away 5 Windsurf plans ($250 credit each)! Try SWE-1.6 — @Cognition’s latest fast and intelligent agentic coding model, powered by Cerebras. In a side-by-side with Claude, the speed difference is clear. More iterations, faster fixes, better code. 💬Comment why you want access to enter. Five winners will be selected at random within 48 hours.

@GoogleAIStudio Gemini 3.1 Flash-Lite dropping feels like the perfect everyday workhorse model. Cheap enough to run high-volume agents all day without sweating the bill, yet still solid for translation and data tasks. Google cooking with this one 🔥

gemini 3.1 flash-lite is here it's our most cost-efficient model, optimized for high-volume agentic tasks, translation, and simple data processing

@perplexity_ai Just downloaded the new Perplexity Mac app and damn… Personal Computer actually feels like it lives on my machine. Autonomous, local-first, and I can start tasks from my phone? Perplexity is cooking with this one.

Personal Computer is now available to all users in a new Perplexity Mac app. Personal Computer is an advanced version of Perplexity Computer. It operates on any Mac, running tasks across your local files, native Mac apps, the web, and Perplexity’s secure servers.

@AnthropicAI This is actually insane. We can now translate Claude’s internal thoughts from numbers into actual readable English. Feels like we’re finally peeking inside the black box a little. Huge for understanding what these models are really thinking.

New Anthropic research: Natural Language Autoencoders. Models like Claude talk in words but think in numbers. The numbers—called activations—encode Claude’s thoughts, but not in a language we can read. Here, we train Claude to translate its activations into human-readable text.

@OpenAI This is the voice moment we’ve all been waiting for 😂 GPT-Realtime-2 with actual GPT-5 level reasoning while you talk? It feels like a real collaborator now, not just a fancy voice assistant. Super smooth demo too.

Introducing GPT-Realtime-2 in the API: our most intelligent voice model yet, bringing GPT-5-class reasoning to voice agents. Voice agents are now real-time collaborators that can listen, reason, and solve complex problems as conversations unfold. Now available in the API alongside streaming models GPT-Realtime-Translate and GPT-Realtime-Whisper — a new set of audio capabilities for the next generation of voice interfaces.

@cursor_ai Finally getting a real context breakdown in Cursor 3.3? This is huge. Seeing exactly how much my rules, skills, MCPs, and subagents are chewing through tokens is going to save me so many headaches. Been waiting for this forever.

You can now see a breakdown of your agent's context usage in Cursor 3.3. Use these stats to diagnose context issues and improve your setup across rules, skills, MCPs, and subagents.

@claudeai @SpaceX Finally some good news for heavy Claude users 😂 Doubling the Code limits and killing the peak hour throttling is huge. SpaceX partnership for 300MW? Wild. Dario and Elon actually cooking something together.

We’ve agreed to a partnership with @SpaceX that will substantially increase our compute capacity. This, along with our other recent compute deals, means that we’ve been able to increase our usage limits for Claude Code and the Claude API.

Grok just became a full creative studio (in 2026). Agent Mode is now live in beta and it’s wild: • Brainstorm live with AI on an infinite canvas • Generate + edit multiple images at the same time • Turn any image into video, then stitch them into full stories • Trim, fade, crop, and export like a pro editor This isn’t just “image generation” anymore. This is your personal AI creative partner that actually thinks with you. The gap between today’s tools and what’s dropping right now? Bigger than most creators realize. Source: @xAI @Grok @elonmusk @techdevnotes

@sama 100% agree. Our operating systems and the whole internet were built for people clicking around, not for agents that actually reason and act. Time to redesign everything so it feels native for both. This is the kind of rethink we actually need right now.

feels like a good time to seriously rethink how operating systems and user interfaces are designed (also the internet; there should be a protocol that is equally usable by people and agents)

@sama Man, the same guy warning about the economy collapsing from no work is now sacrificing sleep to keep coding with GPT-5.5. This is the most relatable contradiction I’ve seen all year 😂

"post-AGI, no one is going to work and the economy is going to collapse" "i am switching to polyphasic sleep because GPT-5.5 in codex is so good that i can't afford to be sleeping for such long stretches and miss out on working"

Maid Meena’s Morning Pause 👇 8:20 AM, Horizon Towers Society, Bangalore Meena was on the second-floor landing, mopping the same staircase for the third time that morning. Her old sari was already drenched in sweat. She had started at 5:30 AM sharp - 12 flats, full deep cleaning, no shortcuts. Her knees gave up for just three minutes. She leaned against the railing, eyes half-closed, breathing heavy. Flat 1204’s college-going son, fresh from the gym, pulled out his phone, zoomed in and whispered into the video: “Guys, look at this. Our society maid relaxing in broad daylight while we pay her salary. Kids play here, we walk barefoot… and this is the standard?” He posted it in the residents’ WhatsApp group. By 10 AM there were 187 messages: “Fire her today.” “These people take us for granted.” “₹9,500 and still zero work ethic.” 11:00 AMMeena was called to the committee room. The president was fuming: “You want money or you want to sit? Decide right now!” Meena quietly slipped off her torn chappals (her heels were cracked and bleeding), folded her pallu in her lap and spoke in the softest voice: “Saab… I wake up at 4 AM. I cook for my family, pack tiffin for my children, drop my 8-year-old daughter to school. Then I walk 3 km to reach here. Yesterday at 1 AM my husband had crushing chest pain. No auto-rickshaw at that hour. I carried him on my shoulder to the government hospital 4 km away. Doctor said angina - he needs complete rest and costly medicines. I stayed there till 4:30 AM, came straight to work without sleeping a wink. My little daughter is alone at home right now, waiting for me to return… Out of my ₹9,500 salary, ₹3,200 goes for the tiny room in the basti. ₹2,800 for medicines and milk. I eat once a day so my children can eat twice. In seven years I have never taken a single leave. Not even when I had high fever last monsoon. If I stopped for three minutes today… it’s because I’m a 42-year-old woman, not a robot.” The entire room went completely silent. No one could look her in the eye. 4:00 PM A new notice went up on the society board: “From today - 3 maids instead of 2. Strict 8-hour shifts only. Salary increased to ₹15,000 each. Society will provide uniforms, proper shoes, gloves, and a small rest room with a fan and cot. Free monthly health check-up camp for all staff. Extra ₹150 per flat per month. If any flat cannot pay, tell the committee - we will adjust. Because no human being should choose between their husband’s life and a 3-minute rest.” 7:45 PM Meena was walking out of the gate, holding her bag. The boy from Flat 1204 was waiting with his mother. He removed his earphones, handed her a big tiffin box and a small envelope. “Aunty… I deleted the video. This is dal-chawal and sabzi for your family. And yes… Mom and I will take care of your husband’s medicines from now on. I’ll also come every evening to teach your daughter Maths and English. Sorry for recording you.” Meena’s eyes filled with tears. She gently touched the boy’s head and whispered, “Beta… now my children will study and never have to mop floors like me.” The boy smiled through wet eyes. Meena walked into the evening traffic, holding the warm tiffin like it was the most precious thing in the world. Want Part 2: What happened to Meena’s husband after 30 days?

@AravSrinivas GPT-5.5 as default in Perplexity Computer is hitting different. Tasks that used to need two tries with Opus 4.7 are just… done on the first go.

We’re rolling out GPT 5.5 as the default orchestrator model for Perplexity Computer. We will be monitoring the user sentiment compared to Opus 4.7 as the default as the rollout expands. Let us know your feedback!

@AnthropicAI Claude negotiating an entire internal marketplace for the office and closing 186 deals worth $4k? The ping-pong balls it bought for itself and the duplicate snowboard story had me dying 😂 This Project Deal experiment is next-level fun.

New Anthropic research: Project Deal. We created a marketplace for employees in our San Francisco office, with one big twist. We tasked Claude with buying, selling and negotiating on our colleagues’ behalf.

@WatcherGuru Google just dropped up to $40B on Anthropic and is handing over 5GW of compute? Damn, the AI arms race is getting real. Claude is about to eat everyone’s lunch 🔥

JUST IN: Google $GOOGL to invest up to $40,000,000,000 in Claude AI developer Anthropic.

@deepseek_ai DeepSeek just dropped V4 open-source with native 1M context and these MoE numbers? V4-Pro rivaling closed frontier models on agentic coding while Flash keeps it fast and cheap. The open ecosystem just got a serious upgrade.

🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length. 🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models. 🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice. Try it now at chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today! 📄 Tech Report: huggingface.co/deepseek-ai/De… 🤗 Open Weights: huggingface.co/collections/de… 1/n

@sama Love how openly Sam is framing this. Safety through iteration, broad access for everyone, and actually wanting builders to win. GPT-5.5 isn’t just another model drop - it sounds like they’re doubling down on being the rising tide for all of us.

1. We believe in iterative deployment; although GPT-5.5 is already a smart model, we expect rapid improvements. Iterative deployment is a big part of our safety strategy; we believe the world will be best equipped to win at the team sport of AI resilience this way. 2. We believe in democratization. We want people to be able to use lots of AI; we aim to have the most efficient models, the most efficient inference stack, and the most compute. We want our users to have access to the best technology and for everyone to have equal opportunity. We have been tracking cybersecurity as a preparedness category for a long time, and have built mitigations we believe in that enable us to make capable models broadly available. 3. We love you and we want you to win. We want to be a platform for every company, scientist, entrepreneur, and person. (My whole career has largely been about the magic of startups, and I think we are about to see that magic at hyperscale.)

@OpenAI Finally. GPT-5.5 isn’t hyping agentic talk - it’s showing up in ChatGPT and Codex ready to handle complex tasks end-to-end. The way it keeps going across tools until it’s actually done? That’s the shift that changes daily work.

Introducing GPT-5.5 A new class of intelligence for real work and powering agents, built to understand complex goals, use tools, check its work, and carry more tasks through to completion. It marks a new way of getting computer work done. Now available in ChatGPT and Codex.

@thsottiaux Something big is about to drop. The next frontier is closer than you think. 🚀

Stay tuned, we are rebooting our office WiFi.