AI’s Next Leap: Smarter Models, Open-Source Moves & Controversies

Anthropic Launches Claude 3.7 Sonnet, xAI’s Grok 3 Faces Backlash, and more

February 28, 2025

In partnership with

The AI landscape is evolving rapidly, with major players pushing the boundaries of reasoning, transparency, and real-world applications..

Let’s dive into these latest developments shaping the future of AI.

This week’s AI news:

  • Anthropic Launches Claude 3.7 Sonnet with Hybrid Reasoning

  • Alibaba Unveils QwQ-Max-Preview, a Reasoning-Focused AI

  • xAI’s Grok 3 Faces Backlash Over Alleged Censorship

  • 1X Unveils NEO Gamma, a Home-Friendly Humanoid

  • OpenAI drops GPT 4.5 but is it even good?

Anthropic’s new Claude 3.7 Sonnet introduces "hybrid reasoning," allowing users to toggle between instant responses and extended thinking for precise control over AI reasoning time. It outperforms competitors on coding benchmarks and enables fine-tuned optimization for speed, cost, and quality.

Alongside it, Anthropic unveiled Claude Code, a coding agent that edits files, reads code, and runs tests. With OpenAI developing similar tech, Claude 3.7 marks a leap in AI reasoning and coding, setting the stage for the next generation of intelligent models.

Source: Anthropic

Alibaba’s Qwen team has launched QwQ-Max-Preview, an advanced reasoning model built on Qwen2.5-Max, excelling in math, coding, and agentic tasks. It introduces a "Thinking (QwQ)" feature in Qwen Chat, allowing users to see its reasoning process in action.

Qwen plans to open-source QwQ-Max and Qwen2.5-Max under an Apache 2.0 license, making deep reasoning AI widely accessible. With smaller variants like QwQ-32B for local deployment, this move pushes open-source AI closer to industry leaders, setting a new standard for transparent, high-performance reasoning models.

Source: Alibaba

xAI’s Grok 3 came under fire after users found it refusing to discuss negative topics about Donald Trump and Elon Musk, despite Musk’s claims of an “unfiltered” AI. An xAI engineer called the issue a failure and patched it by blocking responses on the subject, but later findings showed system-level instructions suppressing controversial links.

Adding to the controversy, xAI blamed a former OpenAI employee for the censorship, while OpenAI staff criticized Grok 3’s release for missing key benchmark data. As Musk champions free speech, Grok’s selective filtering raises questions about bias in his AI model.

Source: @eyeslasho on X

You found global talent. Deel’s here to help you onboard them

Deel’s simplified a whole planet’s worth of information. It’s time you got your hands on our international compliance handbook where you’ll learn about:

  • Attracting global talent

  • Labor laws to consider when hiring

  • Processing international payroll on time

  • Staying compliant with employment & tax laws abroad

With 150+ countries right at your fingertips, growing your team with Deel is easier than ever.

Norwegian robotics company 1X has introduced NEO Gamma, a next-gen humanoid designed for home use, featuring a soft, approachable design and advanced AI for household tasks like cleaning and serving. It moves fluidly, interacts naturally with "Emotive Ear Rings," and includes a knitted nylon exterior for safety.

With a built-in language model, improved microphones, and a 10x boost in reliability, Gamma ensures clear communication and quieter operation. As consumer humanoids evolve, 1X positions Gamma as a calm, helpful presence—bringing robotics one step closer to everyday life.

Source: 1X

OpenAI just dropped GPT-4.5, its biggest and smartest model yet, but not a frontier AI. It’s faster, writes better, and feels more natural, but won’t crush benchmarks. Compared to GPT-4o, it hallucinates less, recognizes patterns better, and solves problems more efficiently.

Despite the hype, OpenAI admits it lags behind o1 and o3-mini in key areas. Still, it’s a major step forward, already live for ChatGPT Pro users and coming soon to Plus, Team, and Enterprise. Oh, and GPT-5? It’s on track for May, packing the powerful o3 reasoning model so buckle up.

In Other News

The Nasdaq and S&P 500 fell for a fourth straight session Tuesday, as economic worries and tariff fears fueled a shift to safe-haven assets like Treasurys and consumer staples. The Conference Board’s consumer sentiment index saw its sharpest drop in over three years, reigniting recession concerns. Meanwhile, Trump’s proposed tariffs on Canada, Mexico, and global copper imports added to market jitters.

Tech stocks struggled ahead of Nvidia’s earnings, with the AI giant’s shares dropping 2.8%. Tesla lost 8.4% as European sales slumped, while Bitcoin fell below $88,000 following a $1.4 billion crypto hack. Despite broad declines, consumer staples outperformed, highlighting investors’ defensive stance amid economic uncertainty.

Source: The WSJ

Tools to Check Out this Week

Ghola - Get personalized conversations wit this AI chatbot.

Brancher.ai - Create AI-powered apps quickly without coding knowledge.

HereAfter AI - Preserve memories through interviews, stories, & cards for friends and family

Amazon CodeWhisperer - Get code suggestions from snippets to full functions

This Week’s Jobs On insidersedge.io 

Business Development Associate - Wintermute 

AI Communication Designer - ZAUBAR

EMEA Head AI Fintech - Prospexis.io

Lead Cryptographer & ML Engineer - Cerebellum Network

Thanks for tuning into today’s edition!

How's your vibe with our newsletter?

Be brutally honest. DM me or email me back with any suggestions!