|
|
Hi there, this is your daily ☕️ Techpresso.
|
|
|
In today's Techpresso:
|
|
📺 YouTube surpasses Disney as largest media company 👀 SpaceX Starship Moon lander faces new delays 🔧 Meta unveils four new in-house AI chips 💥 Asus co-CEO calls MacBook Neo a shock 🛒 Amazon wins court order to block Perplexity's AI shopping agent 🛡️ Meta deploys AI scam detection across all patforms 🎁 + 22 other news & articles you might like 🧰 + 5 trending tools 📚 + 3 trending papers & reports
|
|
FROM OUR PARTNER
Every AI agent demo looks magical, but most hit a wall in enterprise deployment. It's not model quality or latency. It's authorization. Authentication proves an agent's identity. Authorization defines its blast radius. The winners in enterprise AI won't have the most features.They'll be the ones enterprises can safely trust. Learn how WorkOS FGA scopes that blast radius with resource-level permissions. Read the deep dive →
|
📺 YouTube surpasses Disney as largest media company
LINK
|
- YouTube has passed Disney and become the largest media company by ad revenue, pulling in $40.4 billion in 2025 — more than Disney, NBC, Paramount, and Warner Bros. Discovery's combined $37.8 billion.
- This marks a big turnaround from 2024, when YouTube's $36.1 billion in ad revenue fell short of the four major Hollywood studios' collective $41.8 billion, according to research firm MoffettNathanson.
- Parent company Alphabet reported YouTube's total revenue soared to $60 billion in 2025, with a big portion now coming from subscriptions like YouTube TV, YouTube Premium, and NFL Sunday Ticket.
|
👀 SpaceX Starship Moon lander faces new delays
LINK
|
- SpaceX's Starship lunar lander for NASA still faces numerous technical challenges and likely more delays ahead of a planned 2028 moon landing, according to a new government audit from NASA's Office of Inspector General.
- The biggest challenge is demonstrating orbital refueling, a process the OIG called "entirely new" since it has never been done vehicle-to-vehicle, and SpaceX intends to attempt it for the first time this year.
- SpaceX and NASA have also disagreed over whether the Starship design gives astronauts enough manual control, and SpaceX may request a waiver to automate the vehicle in order to stay on schedule.
|
🔧 Meta unveils four new in-house AI chips
LINK
|
- Meta has announced four new generations of its custom MTIA silicon chips, designed to handle ranking, recommendations, and GenAI workloads, with all four set to be developed and deployed within the next two years.
- MTIA 300 is already in production for ranking and recommendations training, while MTIA 400, 450, and 500 will primarily support GenAI inference production through 2027 across Meta's apps.
- Meta says it can release new chip generations every six months or less by reusing modular designs, which is much faster than the typical industry cycle of one to two years.
|
💥 Asus co-CEO calls MacBook Neo a shock
LINK
|
- Asus co-CEO S.Y. Hsu called the $599 MacBook Neo a "shock" to the entire Windows PC industry, saying Apple's move to release a budget-friendly product will force PC makers to respond.
- Despite calling the MacBook Neo a shock, Hsu argued it could have limited appeal due to its non-upgradeable 8GB of unified memory and its role as a "content consumption" device like an iPad.
- Hsu said all PC players including Microsoft, Intel, and AMD take the MacBook Neo threat seriously, and the entire Windows PC ecosystem will push out products to compete against Apple.
|
🛒 Amazon wins court order to block Perplexity's AI shopping agent
LINK
|
- A federal judge issued a preliminary injunction blocking Perplexity's Comet browser from accessing Amazon user accounts or making purchases, marking the first major court ruling on autonomous AI shopping agents.
- The court ruled that a user's permission to an AI agent does not substitute for platform authorization, finding Perplexity likely violated the Computer Fraud and Abuse Act after disguising Comet as Chrome.
- Perplexity has until March 17 to ask the Ninth Circuit to stay the order, and the ruling could set a legal template governing how all autonomous agents access third-party platforms.
|
🛡️ Meta deploys AI scam detection across all patforms
LINK
|
- Meta is rolling out new AI-powered scam detection features across Facebook, WhatsApp, and Messenger designed to alert users before they interact with suspicious accounts or messages.
- Facebook will test alerts warning users about suspicious friend requests, while WhatsApp will now flag potentially fraudulent device linking requests that scammers use to hijack accounts.
- Meta said it removed more than 159 million scam ads last year, with 92 percent taken down before anyone reported them, plus 10.9 million accounts tied to criminal scam centers.
|
|
FROM OUR PARTNER
🗣️Voice Agents need speed you trust Speechmatics' Voice Agent API delivers partials in <250ms and finals around 250-300ms. Built real-time first, not batch models forced into streaming. That's why your voice agents get unmatched accuracy at conversational latency. What you get: - 55+ languages, accent-agnostic performance
- Native LiveKit, Pipecat, Vapi support
- Deployment in the cloud, on-prem, hybrid, on-device
👉 Start building with $200 free credits. |
|
|
|
Other
news & articles you might like
-
Anthropic Tells Judge Billions at Stake If US Shuns AI Tool
LINK
-
Oracle shrugs off debt critics, touts AI "halo" in strong earnings
LINK
-
Amazon launches its healthcare AI assistant on its website and app
LINK
-
Google expands Gemini DoD partnership with Gem-like agents for unclassified projects
LINK
-
ChatGPT can now create interactive visuals to help you understand math and science concepts
LINK
-
Google wraps up $32B acquisition of cloud cybersecurity startup Wiz
LINK
-
YouTube will alert civic leaders and reporters to deepfakes that involve their likeness
LINK
-
Anthropic Launches Internal Think Tank as Pentagon Blacklist Heads to Court
LINK
-
Report: Sora AI Video Generation May Be Coming to ChatGPT
LINK
-
Amazon Seeks Debt To Pay For AI Investments
LINK
-
Polymarket and Palantir team to protect integrity of sports betting as prediction markets face key moment
LINK
-
AI-powered apps struggle with long-term retention, new report shows
LINK
-
Amazon insists AI coding isn't source of outages
LINK
-
Human brain cells set to power two new data centers, thanks to 'body-in-the-box' CL1
LINK
-
After complaints, Google will make it easier to disable gen AI search in Photos
LINK
-
Hisense TVs force owners to watch intrusive ads when switching inputs, visiting the home screen, or even changing channels
LINK
-
Meta didn’t buy Moltbook for bots — it bought into the agentic web
LINK
-
AI should help us produce better code
LINK
-
DOJ Investigating if Iran Used Binance to Evade Sanctions: WSJ
LINK
|
|
🧰 Trending tools
|
InsForge: a backend platform with databases, auth, storage, and APIs designed for AI agents to autonomously build and deploy full-stack applications.
LINK
|
|
Cardboard: an AI-powered video editor that analyzes raw footage and automatically creates polished cuts based on your instructions, eliminating manual timeline editing.
LINK
|
|
Teract AI: automates LinkedIn and X engagement by mimicking your writing style to post comments, replies, and trending content while you focus elsewhere.
LINK
|
|
OpenUI: converts LLM text responses into interactive UI components like charts and forms through a simple API integration with just two lines of code.
LINK
|
|
Gemini Embedding 2: converts text, images, video, audio, and documents into unified embeddings for cross-modal search and classification in a single vector space.
LINK
|
|
You can check the previous tools here, or add your tool here
|
|
📚 Trending papers & reports
|
Bringing AI to Finance: Your free guide to understanding how intelligent automation is transforming finance teams, improving accuracy, and accelerating day-to-day operations. Get the free AI guide.
|
|
> A 72 billion parameter language model was trained across 176 consumer GPUs spread over the internet, achieving performance comparable to centralized training while reducing costs by distributing compute among untrusted participants.
LINK
|
|
> A new number format for AI was designed in one week using LLMs as coding assistants, demonstrating that large language models can accelerate specialized technical work like creating floating point alternatives to bfloat16 and FP8.
LINK
|
|
> Google's quantum computer solved a random circuit sampling problem in 5 minutes that would take classical supercomputers 10,000 years, demonstrating quantum advantage for the first time.
LINK
|
|
|
|
🎓 Want to master the AI tools we cover every day?
Techpresso's AI Academy has 300+ step-by-step tutorials on ChatGPT, Claude, Perplexity, and every tool that matters. No fluff — just practical workflows you can use at work. Try it free for 14 days.
|
|
💬 How did you find today's edition?
We read every reply — just reply to this email and let us know how we can improve!
|
|
|
