|
|
Hi there, this is your daily ☕️ Techpresso.
|
|
|
In today's Techpresso:
|
|
💰 Nvidia keeps the AI party alive 🍌 Google launches Nano Banana 2 🎰 New York sues Valve over loot box gambling 🤖 Perplexity's new AI system can autonomously run projects 🍔 Burger King uses AI to monitor employee politeness 📉 Jane Street accused of manipulating Bitcoin prices 🎁 + 19 other news & articles you might like 🧰 + 6 trending tools 📚 + 5 trending papers & reports
|
|
FROM OUR PARTNER
Join Airia's AI Agent Hackathon ahead of the Australian Race weekend. The Virtual AI Agent Hackathon runs February 23–March 1, giving you seven days to design autonomous AI agents that elevate the fan experience. Top projects will be highlighted ahead of the first race weekend of 2026. Kickoff: Feb 23 | Projects due: March 1, 11:59 PM local time. Ready to build? Register Here:
|
💰 Nvidia keeps the AI party alive
LINK
|
- Nvidia posted record quarterly profits driven by surging AI demand, with CEO Jensen Huang saying token demand has gone "completely exponential" and even six-year-old GPUs in the cloud are fully consumed.
- The company reported $68 billion in quarterly revenue, up 73% from last year, with $62 billion coming from data center sales split between $51 billion in compute and $11 billion in networking.
- Nvidia said it has no China chip export revenue yet despite lifted restrictions, flagged Chinese competitors gaining ground, and confirmed it is close to a partnership agreement with OpenAI.
|
🍌 Google launches Nano Banana 2
LINK
|
- Google has announced Nano Banana 2, an image generation model built on Gemini 3.1 Flash Image that creates more realistic images and will become the default across the Gemini app.
- The model can produce images from 512px to 4K resolution, maintain character consistency for up to five characters, and handle fidelity of up to 14 objects in one workflow.
- All images created with Nano Banana 2 will carry a SynthID watermark and support C2PA Content Credentials, and Google says people have already verified over 20 million images since November.
|
🎰 New York sues Valve over loot box gambling
LINK
|
- New York Attorney General Letitia James has sued Valve Corporation, claiming the loot box systems in Counter-Strike 2, Team Fortress 2, and Dota 2 operate as illegal gambling under state law.
- The complaint says players pay roughly $2.49 for a key to open a virtual case that awards a random cosmetic item, with slot machine-style animations and "near miss" effects mimicking casino techniques.
- Valve profits from a 15% commission on its Steam Community Market, where skins can be resold, and the lawsuit says third-party marketplaces let users cash out items for real money.
|
🤖 Perplexity's new AI system can autonomously run projects
LINK
|
- Perplexity launched Computer, a multiagent orchestration system that routes tasks across 19 frontier AI models to handle full workflows from research and design through code deployment.
- Claude Opus 4.6 serves as the core reasoning engine, breaking down requests into subtasks and assigning each to specialists like Gemini, Grok, or ChatGPT 5.2 based on the task's requirements.
- Available now to Max subscribers at $200 per month, Computer introduces per-token billing for consumers for the first time, making AI budgeting look more like managing cloud compute costs.
|
🍔 Burger King uses AI to monitor employee politeness
LINK
|
- Burger King is rolling out an AI chatbot called "Patty" that lives in employee headsets, helping with meal preparation and evaluating worker interactions with customers for "friendliness."
- The OpenAI-powered system was trained to recognize phrases like "please" and "thank you," and managers can ask the AI assistant how their location is performing on friendliness as a coaching tool.
- Patty is piloting in 500 restaurants now, with the full BK Assistant platform planned for all US locations by the end of 2026, while AI drive-thru ordering is only being tested in fewer than 100 stores.
|
📉 Jane Street accused of manipulating Bitcoin prices
LINK
|
- Jane Street, one of the world's largest trading firms, is facing accusations from crypto traders on X who claim it systematically sold bitcoin at 10 a.m. ET daily to push prices down.
- Data tracked by crypto economist Alex Kruger shows the alleged "10 a.m. dump" closely mirrors Nasdaq performance, suggesting broad risk-asset repricing rather than manipulation by a single firm.
- As an authorized participant in bitcoin ETFs, Jane Street can legally short shares without borrowing costs and source bitcoin privately through OTC shops, which may create temporary downside pressure.
|
|
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 → |
|
|
|
Other
news & articles you might like
-
Anthropic’s Pentagon Showdown Is About More Than AI Guardrails
LINK
-
Instagram to start parent alerts for teen suicide, self-harm searches
LINK
-
Gemini can now automate some multi-step tasks on Android
LINK
-
The White House wants AI companies to cover rate hikes. Most have already said they would.
LINK
-
Anthropic acquires Vercept to advance Claude's computer use capabilities
LINK
-
Alphabet-owned robotics software company Intrinsic joins Google
LINK
-
Kalshi fined a MrBeast editor for insider trading on markets related to the YouTube star
LINK
-
Everything announced at Samsung Unpacked: The Galaxy S26 Ultra, Galaxy Buds 4 and more
LINK
-
Hacker used Anthropic's Claude chatbot to attack multiple government agencies in Mexico
LINK
-
Cloudflare experiment ports most of Next.js API 'in one week' with AI
LINK
-
OpenAI COO says ads will be ‘an iterative process’
LINK
-
Paypal, Stripe not currently in talks
LINK
-
Thrive Capital invested about $1 billion in OpenAI at a $285 billion valuation, source says
LINK
-
'You’ve heard about the SaaSpocalypse? And it isn’t our first. We’ve had a few of them': Salesforce CEO Marc Benioff bullish over AI market potential
LINK
-
Andrej Karpathy says programming is "unrecognizable" now that AI agents actually work
LINK
|
|
🧰 Trending tools
|
Koidex: searches and scans extensions, npm packages, and AI models across multiple platforms to quickly verify if they're safe to install.
LINK
|
|
Rover by rtrvr.ai: automates complex web scraping and browser workflows using natural language commands through multiple interfaces including API, browser extension, and WhatsApp integration.
LINK
|
|
Tessl: a platform for evaluating and optimizing AI agent skills to reduce hallucinations and bugs, helping developers build more reliable agents faster.
LINK
|
|
gpt-realtime-1.5 by OpenAI: enables low-latency voice and audio conversations in applications through real-time streaming API with natural speech interaction capabilities.
LINK
|
|
Terminal Phone: a Bash script enabling anonymous, end-to-end encrypted voice and text walkie-talkie communication over Tor using .onion addresses.
LINK
|
|
IronClaw: a security layer for AI agents that stores credentials in encrypted vaults and sandboxes tools to prevent prompt injection attacks from stealing API keys.
LINK
|
|
You can check the previous tools here, or add your tool here
|
|
📚 Trending papers & reports
|
The habits of high-performing mobile teams: The Bitrise Mobile Insights Report breaks it down using real customer data, showing what drives app development success and what sets high-performing teams apart. Get the full report.
|
|
Large Language Models Can Unmask Anonymous Writers: by comparing writing patterns like word choice and sentence structure, LLMs matched anonymous texts to known authors with 74% accuracy.
LINK
|
|
Combining Multiple LoRA Adapters Degrades Performance: merging several specialized AI model adapters together produces worse results than using a single adapter, even with smart combination methods.
LINK
|
|
Automated theorem prover solves competition problems: the system proved 10 out of 15 mathematical problems from a contest without human help, matching undergraduate level reasoning.
LINK
|
|
Chain of Thought Has Fundamental Limits: even with unlimited thinking steps, language models cannot solve all solvable problems because sequential reasoning hits mathematical boundaries.
LINK
|
|
Hexagon-MLIR Compiler for Qualcomm Neural Processors: software translates AI models into optimized instructions for Qualcomm's specialized chips, letting developers run neural networks on mobile devices efficiently.
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!
|
|
|
