- TheNeural
- Posts
- This Week in AI with The Neural
This Week in AI with The Neural
Robots Run Marathons, AI Spots Cancer, Perplexity Tracks, Microsoft Hires Agents
TECHIES, ASSEMBLE!
AI Highlights
AI - POINT OF VIEW
Can Machines Grasp Our Humanity?
Debate rages whether AI’s pattern mimicry equals understanding or consciousness emerges from quantum gravity. Explore arguments that keep humanity just beyond code’s reach—for now.
AI INDUSTRY
OpenAI Frees Deep Research
A lightweight “deep research” tool quietly rolled out to Free and tiered users, signaling OpenAI’s bet that broad access to insight engines drives mass adoption.
Grok Speaks Your Language
X’s Grok now supports six languages, including Hindi and French, and adds real-time voice search—pushing its AI assistant into global, audio-first interactions.
Meet Microsoft’s Agent Boss
The 2025 Work Trend Index unveils “Agent Boss,” letting employees manage AI teammates. New Copilot features aim to make these ever-on assistants part of your workflow.
Anthropic’s Dark Misuse Revelations
Anthropic disclosed chilling case studies of Claude’s misuse—from social media manipulation to malware writing—prompting new defense layers and detection tools for AI safety.
Perplexity’s Comet: Your Data Spy
Comet, Perplexity’s new browser, will track activity beyond its app to fuel hyper-targeted ads—echoing Google’s playbook and reigniting debates over privacy trade-offs.
Manus AI’s $75 M Power Play
Butterfly Effect’s Manus AI platform just closed a $75 M Series A, valuing the startup at $500 M as it preps U.S., Japan and Middle East expansion.
Kling 2.0’s Cinematic Leap
Kling 2.0’s April rollout brings 1080p video, multi-element editing, physics-driven effects and a Motion Brush for selective movement—plus performance fixes that flex its AI video prowess.
Cursor’s Costly Hallucination
Cursor AI bot falsely claimed a one-device install policy, nearly losing customers before apologizing and labeling AI-generated replies to rebuild trust.
AI AROUND US
ChatGPT’s Life-Saving Hunch
After vague symptoms stalled doctors, ChatGPT suggested tests that revealed Lauren Bannon’s cancer—highlighting AI’s role as a crucial second opinion in complex diagnoses.
AI’s Invisible Surface Ad
Microsoft’s recent Surface ad used generative AI to stitch scenes seamlessly—so subtle viewers missed it, proving hidden efficiency can trump flashy AI.
MIT’s Periodic ML Table
MIT researchers created I-Con, a “periodic table” organizing 20+ classic algorithms to reveal shared mechanics and predict new AI methods waiting to be discovered.
Robots Race and Stumble
Beijing’s half-marathon debuted humanoid runners; only four finished, with Tiangong Ultra trailing the human champ by an hour—spotlighting both strides and stumbles in robo-mobility.
Drones That Chase Lightning
NTT engineers in Japan coax lightning with drones using electric-field control—paving paths for infrastructure defense and even future energy harvesting from nature’s fury.
Image Credit: The Verge
When Radio Host Isn’t Real
Australian station CADA aired “Thy,” an AI-cloned voice for six months without disclosure. The reveal sparked backlash and a debate on listener trust.
‘HOW TO’ WITH AI
Prompting Mastery: Google’s Fast-Track Course
Google’s free Coursera “Prompting Essentials” teaches you to craft effective AI prompts in under 10 hours, covering chaining techniques, everyday automation, and certification to boost your skills.
FROM THE NEURAL - High-Performing AI Culture
The Neural Talks: The Culture Behind High-Impact AI
Girish, CTO at Scribble Data, joins The Neural Talks to break down what makes an AI team truly high-performing. From setting clear north stars to fostering deep ownership, he shares how to scale AI teams that don’t just build—but build right.
Watch now — you can thank us later!