Meet the new Alexa

Amazon is introducing Alexa+, a next-generation AI assistant powered by generative AI, designed to be more conversational, intelligent, and personalized. Alexa+ goes beyond simple voice commands, enabling natural and expansive conversations that feel more like talking to a helpful friend. It can summarize complex topics, assist with daily tasks, provide entertainment, and even handle tasks like making reservations, managing your smart home, and helping with online shopping. Alexa+ can seamlessly integrate with services like OpenTable, Amazon Music, Grubhub, and Ring, making it more capable than ever.

At the core of Alexa+ is an advanced AI architecture leveraging large language models (LLMs) and “experts,” a system that allows Alexa to orchestrate actions across thousands of services and devices. This includes controlling smart home products, arranging repairs, managing schedules, and even making proactive recommendations based on real-time information. Users can also personalize Alexa+ by having it remember key details such as dietary preferences, important dates, and family routines, making interactions more useful and context-aware. Alexa+ will be accessible across multiple devices, including Echo devices, mobile apps, and web browsers, ensuring continuity across different environments.

Privacy and security remain a priority, with a centralized Alexa Privacy Dashboard and AWS-backed infrastructure ensuring user control over data. Alexa+ will be available for $19.99 per month but will be free for all Amazon Prime members as part of their subscription benefits. Rolling out first in the U.S., early access will be prioritized for owners of Echo Show 8, 10, 15, and 21 devices. With its enhanced intelligence, expanded capabilities, and seamless integration, Alexa+ represents the next evolution of AI-powered assistance.

Get More From Every Word

Meet VoiceHub from Rev - the #1 platform for recording, transcribing, and analyzing speech. It’s revolutionizing how businesses handle their most valuable asset: conversations. 

What sets VoiceHub apart? Their AI accuracy beats Microsoft, Google, and other leading providers in enterprise environments.

But it's not just about accuracy—it's about changing the way you use those accurate transcripts. With VoiceHub, you get:

• Universal capture of audio and video across mobile, desktop, and meetings

• Best-in-class AI transcription and summaries in seconds

• Enterprise-grade security with HIPAA, SOC 2 Type II compliance, and SSO

• Seamless integration with 18 major tools like Zoom and Slack

OpenAI expands Deep Research to all paying ChatGPT users

OpenAI has expanded access to its Deep Research tool, making it available to all paying ChatGPT users. Initially launched in early February, Deep Research was previously restricted to Pro subscribers who paid $200 per month. As of today, Plus, Team, Edu, and Enterprise users can now utilize the feature. Plus users will receive 10 Deep Research queries per month as part of their plan, while Pro users will see an increase from 100 to 120 monthly queries.

In addition to the broader rollout, OpenAI has introduced improvements to Deep Research. The tool now embeds images alongside citations, enhancing the depth of insights it provides. It also has a better understanding of file types, which leads to improved document analysis. These upgrades aim to make the feature more effective for generating comprehensive reports on a wide range of subjects.

To use Deep Research, users simply need to enter a prompt as usual and select the Deep Research icon before submitting their request. Depending on the complexity of the query, responses can take between five and 30 minutes to generate due to the tool's high computational demands. OpenAI has stated that, for now, Free-tier users will not have access to Deep Research, as it remains resource-intensive.

Inception emerges from stealth with a new type of AI model

Inception, a Palo Alto-based AI startup founded by Stanford professor Stefano Ermon, has unveiled a new type of AI model called a diffusion-based large language model (DLM). Unlike traditional large language models (LLMs), which generate text sequentially, Inception’s DLMs apply diffusion techniques that allow for generating and refining large blocks of text in parallel. This approach results in significantly faster performance and lower computing costs. Ermon, who has been researching this method for years, achieved a breakthrough with a student, leading to the company’s formation last summer alongside co-founders Aditya Grover and Volodymyr Kuleshov.

Inception claims its models can run up to 10 times faster than traditional LLMs while costing 10 times less. The company has already secured multiple Fortune 100 customers by addressing the growing need for reduced AI latency and increased speed. Ermon explained that DLMs make more efficient use of GPUs, enabling higher performance. The company provides an API, on-premises and edge deployment options, and support for fine-tuning, with models optimized for various tasks, including code generation and question-answering.

A company spokesperson stated that Inception’s small coding model rivals OpenAI’s GPT-4o mini in quality while being significantly faster. Additionally, their mini model outperforms open-source alternatives like Meta’s Llama 3.1 8B, achieving speeds of over 1,000 tokens per second. While Ermon did not disclose funding details, TechCrunch reports that the Mayfield Fund has invested in the startup. If Inception’s claims hold up, its diffusion-based approach could revolutionize how AI language models are built and deployed.

Other stuff

• Meet the journalists training AI models for Meta and OpenAI

• Apple AI tool transcribed the word 'racist' as 'Trump'

• ChatGPT Saved My Life (I’m Writing this from the ER)

• Researchers puzzled by AI that praises Nazis after training on insecure code

Hope you enjoyed today's newsletter

⚡️ Join over 200,000 people using the Superpower ChatGPT extension on Chrome and Firefox.

OR