- Coastal Seven Consulting
- Posts
- 🔵 THE $14B CODE JUST BUILT ITS OWN BOSS.
🔵 THE $14B CODE JUST BUILT ITS OWN BOSS.
A 4-person team at Anthropic just stood back and watched. Claude Code took 10 days to build its own successor and it didn't ask for permission.
LATEST AI BREAKTHROUGHS

AI products stopped being impressive and started being operational.
Claude Code crossed $14 billion in annualised revenue and then used itself to build its own next product in ten days. Perplexity launched an orchestration engine that routes your work across 19 specialist models and runs for days without checking in.
India hosted the India AI Impact Summit 2026, with participation from 118 countries, and secured 88 signatories to the New Delhi Declaration on AI governance, alongside unveiling three sovereign AI models.
Claude Code Hits $14B and Builds Its Own Successor
Anthropic launched Claude Code as a terminal agent, no IDE, no chat window, just a developer and Claude doing the actual work. One year later it's the company's biggest revenue driver.
What it does today:
Reads your entire codebase, edits files, runs commands and commits, in your terminal, VS Code, JetBrains, desktop app, or browser.
Spawn multiple agents that work on different parts of a task simultaneously, one lead agent coordinates, assigns subtasks, and merges results.
Mention @Claude in Slack with a bug report and get a pull request back.
Pipe logs into it, run it in CI, chain it with other tools, it follows Unix philosophy, so it fits into whatever you already use.
CLAUDE.md files carry your coding standards, architecture decisions, and review checklists into every session automatically.
Sessions move across surfaces, start a task on web, pull it into your terminal, hand it off to the desktop app for visual diff review.
A four-person Anthropic team used Claude Code to build Claude Cowork, its own desktop successor, in 10 days. Claude wrote all the code.
Why it matters: A year ago this was a research preview. Today it writes tests, fixes bugs, opens pull requests, reviews code in CI, and answers Slack messages. Engineers who use it aren't moving faster — they're working differently.
Bharat AI Impact Summit 2026 grabs spotlight
New Delhi’s Bharat Mandapam became the center of global AI dialogue from 16 to 20 February at the India AI Impact Summit 2026. Led by the Ministry of Electronics and Information Technology (MeitY) under the IndiaAI Mission, the summit moved beyond policy talk to real deployment. The India AI Expo spotlighted Indian-built solutions across agriculture, healthcare, climate, education, and accessibility, with homegrown startups such as Sarvam AI, Krutrim, and other emerging AI infrastructure and application players showcasing sovereign models and sector-focused platforms.
Alongside the expo, flagship Casebooks across six priority sectors were launched, a Research Symposium connected academia with industry, CEO roundtables mapped investment pathways, and GPAI meetings aligned multilateral governance culminating in a Leaders’ Declaration. At the Formal Opening Ceremony, Narendra Modi unveiled MANAV, India’s human-first AI framework rooted in ethics, accountability, sovereignty, and inclusion.
On investment: Google committed $15B, Reliance $110B over seven years, Adani announced $100B by 2035. Microsoft committed $50B across the Global South.
Why it matters: India isn't positioning itself as an AI market anymore. It's positioning itself as a builder of foundational models, public infrastructure, and governance frameworks. 118 countries in the room and a Leaders' Declaration signed by 88 of them means the next phase of global AI policy won't be written without India at the table.
Perplexity Computer: AI Becomes the Computer
Perplexity CEO Aravind Srinivas spent two months quiet. On February 25 he explained why. His framing was historical. The word computer once meant a person hired to do calculations. Then it became a machine. Srinivas argues it's evolving again into AI that orchestrates a file system, a browser, and a set of tools, running work on the cloud while you sleep. Perplexity Computer is built on that idea.
You describe an outcome. Computer breaks it into tasks, assigns each one to whichever of its 19 models handles that type of work best. Opus 4.6 for reasoning, Gemini for research, Grok for speed, ChatGPT 5.2 for long-context recall, Veo 3.1 for video and runs everything in an isolated environment. Nothing touches your live files.
Samsung Galaxy S26 — Perplexity is built into the OS. Say Hey Plex or hold the side button and it works across Notes, Calendar, Gallery, and Reminders without switching apps. First time Samsung gave a third-party AI that level of system access.
Comet voice mode upgraded — faster, more natural, works across the browser session
New search embeddings — the underlying tech that determines how accurately Perplexity retrieves and ranks information, now state of the art
Why it matters: One week. A multi-agent workflow engine, a flagship phone integration, a voice upgrade, and a core search improvement. That's not a product cycle. That's a company moving at a different speed.
TOP 5 AI INSIGHTS YOU SHOULD KNOW
OpenAI Signs Pentagon Deal Hours After Anthropic Gets Banned Trump banned Anthropic over two refusals: no mass surveillance, no autonomous weapons. Hours later OpenAI signed a similar deal under the same stated restrictions. The Pentagon accepted. Anthropic is challenging the ban in court.
Google Drops Nano Banana 2 and its free for Everyone Pro-level image generation is now free in 141 countries. Faster, more accurate, generates up to 4K. Default across Gemini, Search, and Flow. Previously behind a paywall.
Nvidia reported $68B in Quarterly Revenue up 73% year over year. Data center alone hit $62.3B. Next-gen Vera Rubin chip ships second half of 2026, 10x more performance per watt than Blackwell. Meta signed a multi-year deal to fill its data centers with both.
DeepSeek Trained on Banned US Chips A Trump official confirmed DeepSeek's V4 was trained on Nvidia's Blackwell chips, banned from export to China. Huawei got weeks of early access. Nvidia and AMD got nothing. DeepSeek reportedly plans to claim it used Huawei chips publicly.
Figma Connects Directly to OpenAI Codex Generate code from a Figma design, turn code back into an editable file without switching tools. Codex usage up 400% since January. A Claude Code integration launched one week earlier.
TRENDING AI TOOLS YOU CANNOT MISS
SnapPod AI - Turns research, notes and ideas into multilingual podcasts for teams, creators and businesses.
Wispr Flow — Speak naturally, it removes filler words and drops polished text wherever your cursor is. Now on Android.
Google Antigravity — Describe what you want built, agents plan, code, test, and deploy it. Free in public preview.
Emergent — Describe an app, get a deployed full-stack product with frontend, backend, and hosting. No code needed.
Perplexity Comet — AI-native browser that searches, summarises, and acts inside web pages. Upgraded voice mode now live.
WE ARE HIRING
Coastal Seven Technologies — Mangalagiri, Andhra Pradesh
Senior AI / Backend Engineer — Python, LLMs, TensorFlow/PyTorch, FastAPI, Cloud ML
Senior Full Stack Engineer — Python, React.js, SQL/NoSQL, Docker, AWS/Azure/GCP
Senior Backend Engineer — Python, Microservices, FastAPI/Django, Redis, Kubernetes - Role posting coming soon. Share your resume and be ahead of the queue.
Send your resume to [email protected]
Tritech Infosystems — Hyderabad/Hybrid
Send your resume to [email protected]

