Top Stories

DirecTV Leans into Streaming, Bringing It Under Main Banner

DirecTV is leaning even further into streaming despite its current decision to ditch its “Stream” branding. As of Sunday, the DirecTV Stream landing page had disappeared, its traffic rerouted to DirecTV.com.  The emphasis is now on streaming on any device, “satellite-free and with no annual contract or hidden fees.” Choosing what to watch and how to watch it “should be simple,” which is why DirecTV says it’s now marketing under a single brand. The main messaging on the new landing page includes a build-your-own channel lineup option and streaming via DirecTV’s bespoke dongle Gemini Air. Read more

TikTok Challenger Neptune Lets Users Hide Likes, Followers

Neptune is a new social video app aiming to compete in the short-form space against stalwarts like TikTok, YouTube Shorts and Instagram Reels. Currently in beta with a test base of about 1,000 users, Neptune has amassed a waitlist of 400,000, according to the company. The app’s noteworthy features include the ability to hide likes and follower counts, in furtherance of a desire to prioritize creativity over competition. That goal stems from founder Ashley Darling’s past experience working as a talent director at the beauty and wellness agency OPTYX, where her focus was “underestimated influencers.” Read more

OpenAI’s Affordable GPT-4.1 Models Place Focus on Coding

OpenAI has launched a new series of multimodal models dubbed GPT-4.1 that represent what the company says is a leap in small model performance, including longer context windows and improvements in coding and instruction following. Geared to developers and available exclusively via API (not through ChatGPT), the 4.1 series comes in three variations: in addition to the flagship GPT‑4.1, GPT‑4.1 mini and GPT‑4.1 nano, OpenAI’s first nano model. Unlike Web-connected models (which have “retrieval-augmented generation,” or RAG) and can access up-to-date information, they are static knowledge models. Read more

Netflix Tests Content Recommendations Powered by OpenAI

Netflix is testing a new recommendation engine that uses OpenAI technology to suggest viewing options based on input that goes beyond the usual parameters of cast and genre. The system is being introduced gradually and is already available in Australia and New Zealand where subscribers must opt-in to try it out, reports say, noting it allows input of more nuanced parameters, including mood, to populate search results. The partnership underscores OpenAI’s efforts to have its technology applied practically and commercially as it seeks to transition from a non-profit to a for-profit public benefit business structure. Read more

Researchers Debut Preview of DeepCoder Reasoning Model

A new open-source code reasoning model called DeepCoder-14B-Preview has hit the market. Built atop DeepSeek-R1 and Qwen2.5 using reinforcement learning (RL), it aims to provide more flexibility by combining high-performance code generation with reasoning capabilities for real-world applications. Its performance is said to be comparable to OpenAI’s o3-mini, “but with a smaller footprint,” say its developers, the research-driven AI companies Together AI and Agentica. “We democratize the recipe for training a small model into a strong competitive coder,” explains Together AI. Read more

Also Noted