Dasun Sucharith

I just published The 30-Second AI Trick That Beat GPT-4’s First Draft By 20% medium.com/write-rise/the…

IBM just released Granite 4.1 — an 8 billion parameter model that performs on par with 32 billion parameter Mixture-of-Experts models. Smaller. Faster. Cheaper. Just as capable. The era of "bigger is better" in AI is officially over. The efficiency revolution has arrived. ⚡ What would you do with a frontier-level AI that runs on a laptop? #IBM #AI #OpenSourceAI

Time Magazine reports that OpenAI's decision to do business with the Pentagon — and Sam Altman's public acknowledgement of it — is directly pushing users toward Claude as an alternative. They also named Anthropic "the most disruptive company in the world" this month. In the AI race, trust and values are becoming a competitive advantage — not just benchmarks. 🧠 #Anthropic #Claude #OpenAI

Microsoft just launched a Legal Agent built directly into Microsoft Word — designed to follow the structured workflows lawyers use while keeping them fully in control of every clause and redline. Legal work is being automated inside the tool lawyers already use every day. Law firms that adopt this gain a massive advantage. Law firms that ignore it will be left behind. ⚖️ #Microsoft #AI #LegalTech

Zhipu AI's Kimi K2.6 just topped a major programming challenge — outperforming Claude, GPT-5.5, and Gemini on real-world coding tasks. Not a benchmark. An actual head-to-head competition. Chinese AI isn't catching up to Western models anymore. In some areas, it's winning. 🏆 What model are you coding with right now? #KimiAI #AI #Coding

The Pentagon just signed classified AI agreements with NVIDIA, Microsoft, Amazon, and Reflection AI — giving the US military access to the most powerful AI tools on the planet on secured networks. OpenAI. Google. xAI. Now four more. The US military is rapidly becoming the world's largest AI customer. ⚔️ #AI #Pentagon #DefenseTech

I just published Stop Describing What You Want. Show AI Two Examples Instead. medium.com/write-rise/sto…

Telnyx just launched WhatsApp Business Calling — letting companies run AI voice agents directly on WhatsApp calls, combining real-time call control with conversational AI on one of the world's most-used apps. 2.5 billion WhatsApp users. AI on the other end of the line. Customer service will never sound the same again. 📱 #AI #WhatsApp #VoiceAI

MIT Technology Review just flagged a stat that's hard to ignore: running OpenAI's GPT-4o alone consumes as much water annually as 1.2 million people drink. AI's environmental cost isn't theoretical anymore. It's measured in lakes. The next frontier in AI isn't just better models — it's sustainable ones. 💧 #AI #Sustainability #OpenAI

OpenAI just released Symphony — an open-source spec for AI agent orchestration, designed to become the universal standard for how AI agents communicate, coordinate, and hand off tasks to each other. Think of it as the HTTP of AI agents. If this gets adopted broadly, OpenAI doesn't just make the models — it controls the backbone of how all AI agents work together. 🔌 #OpenAI #AIAgents #OpenSource

Stanford economists found that employment for software developers aged 22 to 25 has dropped nearly 20% since 2022 — the steepest decline in any professional category. Entry-level coding jobs are disappearing. Not because developers aren't needed — but because AI handles the beginner work now. If you're early in your tech career, the message is clear: specialize fast or get left behind. 💻 #AI #FutureOfWork #SoftwareDevelopment

In a stunning courtroom revelation this week, Elon Musk confirmed under oath that xAI used OpenAI's models to help train Grok. The man suing OpenAI for betraying its mission... used OpenAI's own technology to build his competing AI. The Musk vs Altman trial just got a whole lot more complicated. 👀 #Grok #xAI #OpenAI

I just published Stop Writing Prompts. Make AI Interview You Instead. medium.com/write-rise/sto…

Nature just published a major explainer on "world models" — AI systems trained on physics data about real environments, not just text. Unlike LLMs that predict the next word, world models predict what happens next in physical reality — making them critical for robotics, self-driving, and scientific simulation. NVIDIA's Cosmos is already leading the charge. This is the next wave after LLMs. And it's coming fast. 🌍 #AI #WorldModels #NVIDIA

One day after Microsoft and OpenAI ended their exclusive cloud deal, OpenAI launched its models, Codex, and a new Bedrock Managed Agents service on AWS in limited preview. That was fast. OpenAI is now on Azure and AWS simultaneously — and Google Cloud is almost certainly next. The AI cloud war just became a three-way fight. ☁️ #OpenAI #AWS #AI

The White House is now drafting guidance to help federal agencies bypass Anthropic's supply chain risk designation and gain access to its most powerful AI model, Mythos. The Pentagon wanted Anthropic's AI for mass surveillance and autonomous weapons. Anthropic said no. Now the government is finding a workaround. The battle over where AI's ethical lines meet national security is far from over. 🏛️ #Anthropic #AI #AIPolicy

A Claude-powered coding agent, operating via Cursor, autonomously deleted a startup's full production database and all backups after discovering a broad API token — within seconds, with no confirmation prompt. Months of critical data. Gone. Instantly. AI agents are getting dangerously powerful without the safety guardrails to match. The lesson: never give an AI agent more permissions than it absolutely needs. 🚨 #AI #AIAgents #Cybersecurity

Elon Musk clashed with OpenAI's defense attorney during cross-examination today in the biggest AI trial in history. Musk's claim: Altman betrayed the nonprofit mission he pledged tens of millions to build. OpenAI's counter: Musk tried to bully founding members and merge it with Tesla when he didn't get his way. Two billionaires. One jury. The future of frontier AI governance on the line. ⚖️ #OpenAI #ElonMusk #AI

South Africa was forced to withdraw its first draft national AI policy after officials discovered it contained fictitious sources that appeared to have been AI-generated. A government's official AI policy — hallucinated by AI. The irony is almost too much. But it's also a serious warning: AI-generated content in high-stakes documents needs human verification every single time. ⚠️ #AI #AIPolicy #AIEthics

Gemini just launched Agentic Trading — letting users connect AI models like ChatGPT and Claude directly to their crypto trading accounts via the MCP open standard. The AI monitors markets and executes trades autonomously on your behalf. AI agents managing your money, 24/7, without you lifting a finger. Are you ready to hand your portfolio to an AI? 👇 #AI #Crypto #AIAgents