🤖 AI REVIEW FEBRUARY 14, 2025

^{Read Online | Sign Up | Advertise}

Good morning, AI enthusiasts. While other AI labs have been rushing out model after model, Anthropic has been eerily quiet since the launch of 3.5 Sonnet in June of 2024.

But the startup’s long silence is reportedly about to be broken — with a new hybrid reasoning system launching in the ‘coming weeks’ that could be right on time for the industry’s shifting model meta.

 

In today’s AI rundown:

• Anthropic prepares next major Claude model
• YouTube brings AI video generator to Shorts
• Using ChatGPT as your thesis-writing assistant
• Gemini Flash 2.0 leads new AI agent leaderboard
• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

ANTHROPIC

 Anthropic prepares next major Claude model

Image source: Anthropic

The Rundown: Anthropic is set to release a new AI model in the coming weeks, according to a new report from The Information — which will combine traditional language capabilities with advanced reasoning features, excel at coding, and offer developers more control over balancing speed and compute.

The details:

• The hybrid approach will allow the new model to function as either a standard LLM or a deep reasoning engine, adapting to different use cases on demand.
• The model will also introduce a sliding scale system that lets developers precisely control how much reasoning power to allocate to each query.
• At maximum reasoning, the model reportedly shows particular strength in real-world programming tasks and can handle large-scale codebases.
• Recent rumors had suggested that Anthropic already internally had a model better than OpenAI’s o3, but it hadn’t been released due to safety concerns.

Why it matters: While OpenAI, Google, and others have continued rolling out models, Anthropic has been eerily quiet since Sonnet 3.5. A major upgrade could thrust the company right back into the spotlight — and with ChatGPT now shifting to a more hybrid model approach, Anthropic could be well prepared for a potential AI ‘meta’ shift.

TOGETHER WITH TANKA

 Where collaboration meets long-term memory

The Rundown: Tanka is the world’s first B2B universal chat app with AI long-term memory, inspired by decade-long neuroscience research and built on proprietary memory knowledge graph — saving you hours of writing messages and emails with context-aware AI smart replies.

With Tanka, you’ll get:

• A unified communication hub integrated with tools like WhatsApp, Slack, Gmail, Outlook, Telegram, and more
• An AI work assistant for context-rich smart replies, reminders, and actionable insights
• A personalized memory bank that remembers key discussions, achievements, and todos

Click here to get early access to Tanka.

YOUTUBE

 YouTube brings AI video generator to Shorts

Image source: YouTube

The Rundown: YouTube announced that it is rolling out Veo 2, Google DeepMind’s latest video generation model, into its Shorts platform — allowing creators to generate custom video clips and backgrounds directly from text descriptions.

The details:

• Creators can generate video clips or dynamic backgrounds for Shorts with text prompts and can specify styles, camera effects, and cinematic looks.
• The update enhances the existing Dream Screen feature with faster generation times and improved physics for more realistic movement and scenes.
• All AI-generated content will include Google’s SynthID watermarks and clear labeling to maintain transparency about artificial content.
• The feature is launching first in the U.S., Canada, Australia, and New Zealand through the Shorts camera interface.

Why it matters: This update injects state-of-the-art AI video directly into the workflows of content creators across YouTube, taking a giant leap from just backgrounds to full clips and scenes. While this unlocks new creative possibilities, it will likely blur the already fuzzy lines between real and AI content even further.

AI TRAINING

 Using ChatGPT as your thesis-writing assistant

The Rundown: In this tutorial, you’ll learn how to use ChatGPT’s new Deep Research feature to analyze hundreds of academic sources to create comprehensive literature reviews at a doctoral level.

Step-by-step:

1. Access ChatGPT’s Deep Research feature by selecting the “Deep Research” button.
2. Write your query using this template: “Provide a literature survey on [Topic] from [Year], including [Requirements].”
3. Wait 5-30 minutes while it analyzes hundreds of sources and creates citations.
4. Review your report with clickable citations and verified academic sources.

Pro tip: We did an extensive workshop showing prompting tips for Deep Research, open-source alternatives that do not require a subscription, and how to create your own for free here.

PRESENTED BY ARCEE

 Fall for agentic AI that just works

The Rundown: Arcee Orchestra provides a seamless, turn-key platform that simplifies agentic enterprise AI — with purpose-built SLMs to do exactly what users need.

Join Arcee AI for a webinar on Feb. 26th to:

• Discover the impact of our SLMs that power Arcee Orchestra
• Create production-ready AI agents in minutes with our no-code interface
• Explore the benefits of an all-in-one agentic AI solution
• Learn how our intelligent routing delivers precise results for any task

Sign up and let’s make it a date!

AI LEADERBOARDS

 Gemini Flash 2.0 leads new AI agent leaderboard

Image source: Galileo

The Rundown: Galileo Labs just launched a new AI agent leaderboard to rank how effectively models can handle real-world tool interactions, with Google’s Gemini 2.0 and OpenAI’s GPT-4o debuting at the top spots in the evaluation.

The details:

• The leaderboard evaluated 17 top LLMs on 14 benchmarks, including tests on tool usage and selection, long context, complex interactions, and more.
• Flash 2.0 led with a 0.938 score, outperforming more expensive competitors while excelling across the board on benchmarks.
• Open-source models are closing the gap, with Mistral’s latest Small release achieving scores comparable to some premium offerings at lower price points.
• DeepSeek’s V3 and R1 models were absent from the testing due to a lack of function calling support but will be included if the capabilities are added.

Why it matters: Agents are still in the early stages, but the sector is a key focus in 2025, and measuring these capabilities will likely be a major scoreboard going forward. With a range of different strengths and agentic areas, the tests show that choosing an agent may come down to a user’s specific use case over brand names or pricing tiers.

QUICK HITS

 Trending AI Tools

•  Firefly Video Model – Adobe’s IP-safe AI video generation tool
•  Loveable Visual Edits – Fast, precise AI app building with a Figma-like experience
•  Beatoven AI – AI composer for crafting the perfect background music
•  Alice.tech – Turn generic course materials into personalized learning and exam prep

 AI Job Opportunities

•  Scale AI – SEAL Research Scientist, Agent Robustness
•  LabelBox – Technical Program Manager
•  Celestial AI – Thermal Engineer
•  UiPath – Customer Success Manager

 Everything else in AI today

IMAGINE AI LIVE ’25 connects business leaders with AI innovators for three days of enterprise AI strategies at the Fontainebleau Las Vegas May 28-30. Apply today to save 25%.*

Elon Musk revealed that xAI’s Grok 3 would be released in a week or two during a conversation at the 2025 World Governments Summit, saying the model surpasses all others in reasoning capabilities.

Over a dozen major news publishers filed a lawsuit against Cohere, alleging copyright infringement and trademark violations for using their content to train AI models and generating articles that mimicked their brands.

Baidu plans to make its Ernie chatbot and advanced search freely available starting April 1, aiming to boost adoption in the wake of growing competition from DeepSeek.

Apptronik announced a $350M Series A funding round, with plans to scale production of its Apollo robot and expand into healthcare and consumer markets.

Elon Musk’s letter of intent to acquire OpenAI was revealed, with a deadline of May 10, an all-cash offer of $97.4B, and requirements like full access to company records.

*Sponsored Listing

COMMUNITY

 Watch our last exclusive workshop

We just did a full workshop on building your own AI-powered personal assistant to streamline your daily tasks with Dr. Alvaro Cintas, The Rundown’s AI professor.

Watch it here. Not a member yet? Join The Rundown University on a 14-day free trial.

 Share The Rundown, get rewards

We’ll always keep this newsletter 100% free. To support our work, consider sharing The Rundown with your friends, and we’ll send you more free goodies.