Good morning! This is your daily ☕️ Techpresso.

In today's Techpresso:

💥 Amazon gears up to release next-gen Alexa

🚀 Google launches new AI models

🧠 Researchers trained an OpenAI rival in 30 minutes for under $50

🤖 Tesla boosts hiring for Optimus robot production

🔍 Google ends diversity hiring targets

🎁 + 12 other news you might like

🔮 + 4 handpicked research papers and tools

💥 Amazon gears up to release next-gen Alexa LINK

Amazon is preparing to unveil the next-generation Alexa, which may function as an autonomous AI agent, at a product launch event in New York City on February 26.

The updated Alexa is expected to have improved natural language understanding and could perform tasks autonomously, potentially learning routines and managing smart home devices without direct user input.

While the advanced AI features may initially be free with limited usage, Amazon is considering a monthly charge of $5-$10, with the classic version of Alexa remaining free.

🚀 Google launches new AI models LINK

Google has introduced new AI models, including Gemini 2.0 Flash, Gemini 2.0 Pro Experimental, and Flash-Lite, to enhance efficiency and capability in various applications.

The Gemini 2.0 Pro model features a 2-million token context window, enabling it to process about 1.5 million words simultaneously, making it suitable for complex tasks and coding.

Flash-Lite provides a low-cost AI option with improved performance from the Gemini 2.0 series, offering a cost-effective solution for text, image, and video inputs compared to previous models.

🧠 Researchers trained an OpenAI rival in 30 minutes for under $50 LINK

Researchers from Stanford and the University of Washington trained an AI model, s1, in just 30 minutes for under $50 using cloud computing resources.

s1, based on a model from Alibaba's Qwen, was fine-tuned using a distillation process from Google's Gemini 2.0, resulting in performance comparable to top reasoning models like OpenAI’s o1.

The researchers used a dataset of 1,000 curated questions and instructed s1 to "wait" during reasoning, which improved its accuracy, and they shared the model's details on GitHub.

🤖 Tesla boosts hiring for Optimus robot production LINK

Tesla is actively increasing its recruitment to support the large-scale production of the Optimus humanoid robot at its Fremont, California facility, aiming to make it a commercially available product.

The company has posted numerous job listings for roles like Manufacturing Engineering Technician and Production Supervisor to enhance its manufacturing capabilities for the Tesla Bot.

Elon Musk is also focused on hiring skilled software engineers for his Everything app project, emphasizing coding skills over formal educational background or work history with prestigious firms.

🔍 Google ends diversity hiring targets LINK

Google has decided to end its diversity hiring targets and is set to review its diversity, equity, and inclusion (DEI) programs, as stated in an email viewed by the Wall Street Journal.

The tech giant is now among several companies, including Amazon and Meta, that have recently reduced their DEI initiatives amidst criticism from conservative factions.

Google achieved its goal of increasing hires from underrepresented groups by 30% by 2022, but it is now reconsidering the continuation of its diversity reports and the compliance of its programs with new regulations.

Other news you might like

• Google pulls incorrect Gouda stat from its AI Super Bowl ad LINK

• Huawei manages to pull $118 billion in revenue despite strict U.S. sanctions and restrictions LINK

• NASA has entered the chat: Space agency to Twitch stream from orbit LINK

• Quantum teleportation used to distribute a calculation LINK

• OpenAI co-founder John Schulman's brief stint at Anthropic comes to an end LINK

• Lyft uses Anthropic's Claude chatbot to handle user complaints LINK

• The biggest breach of US government data is under way LINK

• ChatGPT search is now available to all users, even those without an account LINK

• Hugging Face clones OpenAI’s Deep Research in 24 hours LINK

• China is quietly pushing ahead with massive 50,000Mbps broadband rollout to leapfrog rest of the world on internet speeds LINK

• AMD promises “mainstream” 4K gaming with next-gen GPUs as current-gen GPU sales tank LINK

Latest research and tools

GITHUB • 173 points View Source
It's unlikely that there will be any further releases of mt32-pi: mt32-pi: a baremetal MIDI synthesizer for Raspberry Pi that emulates classic sound modules used in vintage games, supporting various SoundFonts and offering low latency audio without needing a traditional operating system.

GITHUB • 28 points View Source
The art of engineering team focus: less is more: The art of engineering team focus: a strategy to increase productivity by prioritizing fewer but more impactful tasks to ensure timely delivery and value creation, rather than spreading resources thin across multiple projects simultaneously.

GITHUB • 23 points View Source
Implementation of a RingBuffer in Java with optional FIFO like semantics: Java RingBuffer: a data structure in Java enabling efficient circular storage that supports optional FIFO operations, allowing retrieval from oldest to newest or from the current start, with additional features like state change notifications and entry management.

GITHUB • 20 points View Source
RustDesk – open-source alternative to TeamViewer: RustDesk: an open-source remote desktop software that allows secure connections without configuration, providing full control over data.

Want to get the latest news differently? Find us on:

See you tomorrow for a new dose of ☕️ Techpresso!