It’s Thursday
Here’s what’s worth your time today.
Meta plans to deploy four new generations of its custom AI chips by 2027 as it tries to reduce reliance on external suppliers and lower the cost of running AI systems. The chips, called MTIA 300, 400, 450, and 500, are part of the company’s in-house accelerator program designed to handle tasks like recommendations, ranking systems, and generative AI workloads.
The move reflects Meta’s dual strategy. It will continue spending billions on hardware from Nvidia and Advanced Micro Devices while also building specialized silicon tailored to its own platforms. As AI demand surges across Meta’s products, the company is racing to control more of the infrastructure stack behind its models and services.
Supported by Lightfield
Logging calls. Updating fields. Writing follow-ups you should’ve sent yesterday.
For most founders, it’s 5-7 hours. That’s half a working day, every week, spent on a tool that’s supposed to help you sell, not give you homework.
Lightfield gives you those hours back. Every email, call, and meeting is captured automatically. Follow-ups surface before you forget. And when you need answers — “Why are deals stalling at the proposal stage?” — you ask, and it tells you with proof from real conversations.
Setup takes 4 minutes. Connect your email and calendar.
Use code CTRLALTDEL13 for 3 free months
IBM’s X-Force team says a ransomware group deployed a likely AI-generated malware framework called Slopoly during a recent attack. The script itself wasn’t especially advanced, but it functioned as a working backdoor that helped attackers keep access to the infected server for more than a week. Researchers believe the code shows signs of being produced by a large language model, with extensive comments and structure typical of AI-assisted development.
The bigger takeaway isn’t the sophistication of the malware, but the speed it enables. AI tools can help attackers create new malware frameworks far faster than before, lowering the barrier to entry and allowing cybercriminals to generate disposable tools on demand. Security researchers say this shift could make attacks easier to launch and harder to track, even if the underlying techniques remain relatively simple.
AI coding platform Replit has raised $400 million in a new funding round that values the company at $9 billion, just six months after hitting a $3 billion valuation. The round was led by Georgian Partners with participation from investors including Andreessen Horowitz and Coatue Management. Founder and CEO Amjad Masad also unveiled Agent 4, an AI system designed to help users build apps, websites, and even the foundations of a startup from scratch.
Replit says more than 50 million people now use the platform, and the company hopes to reach $1 billion in annual recurring revenue by the end of the year. The broader trend reflects the rise of “vibe coding,” where AI handles much of the technical work while founders focus on ideas, design, and product direction.
AI app-building platform Lovable says it added $100 million in revenue in February alone, pushing the company past $400 million in annual recurring revenue. The three-year-old startup is part of the “vibe coding” wave, where users build apps and websites using natural language instead of traditional coding. Lovable now has around 8 million users and enterprise customers including Klarna and HubSpot.
Despite competition from AI coding tools developed by major labs, Lovable’s growth has accelerated rapidly, jumping from $100 million ARR last July to $400 million today. The company also stands out for its efficiency, reaching that scale with just 146 employees, giving it one of the highest revenue-per-employee ratios in the software industry.
Netflix has acquired AI film-editing startup InterPositive, co-founded by Ben Affleck, in a deal that could be worth up to $600 million, according to Bloomberg. The tools focus on post-production tasks such as fixing continuity issues and enhancing scenes rather than generating new footage.
The acquisition fits Netflix’s broader push to integrate AI into production workflows as studios look for ways to reduce costs and speed up editing. Rivals like Amazon and Disney are also investing in AI for film and TV. But the shift is already raising concerns across the industry about job displacement and how creators are compensated when AI tools are trained on existing content.