Entering the Age of Algorithmic Power?
2nd March 2026 | Superintelligence Newsletter
Hey Superintelligence Fam 👋
AI is no longer just productivity tech. From defense-linked restrictions to job market disruption, its influence now touches warfare strategy and workforce stability alike. Are we entering the age of algorithmic power?
Breakthroughs in reasoning and agents promise efficiency, yet parallel conversations around military deployment and automation-driven job shifts are growing louder. Can society harness intelligence without amplifying its unintended consequences?
Let’s dive into what’s new this week..
US Government Halts Use of Anthropic AI : A major policy shift signals rising geopolitical tension around AI governance, as federal agencies are directed to stop using Anthropic systems, highlighting growing national security scrutiny over advanced AI deployments.
OpenAI’s $110 Billion Funding Round : OpenAI’s historic $110 billion raise marks the largest private tech funding ever, backed by Amazon, Nvidia and SoftBank to scale infrastructure and accelerate global deployment of frontier AI technologies.
Jamie Dimon Warns of AI Job Disruption : JPMorgan’s CEO urges governments and businesses to prepare early for workforce disruption, noting AI will steadily integrate into operations and reshape employment patterns rather than triggering sudden mass layoffs.
Apple’s AI Wearables Push : Apple is accelerating its shift into AI-powered hardware, positioning visual intelligence as the core capability behind upcoming wearable devices expected to redefine how users interact with personal computing ecosystems.
Nano Banana 2 : Nano Banana 2 transforms ideas into vivid visuals using conversational prompts, enabling fast image creation, editing, and consistent subject rendering through Gemini’s deep multimodal understanding.
Gemini 3.1 Pro GA : Gemini 3.1 Pro is built for complex reasoning, helping developers and enterprises synthesize multimodal data, solve difficult problems, and build advanced applications with improved thinking and contextual intelligence.
Claude Plugins : Claude Plugins extend AI capabilities by adding custom skills, agents, hooks, and integrations, enabling teams to automate workflows and build reusable intelligence across projects and development environments.
Deep-Thinking Tokens : Google introduces Deep-Thinking Ratio, showing reasoning quality correlates positively with accuracy (r = 0.683), unlike token length (r = -0.59), while Think@n boosts AIME accuracy to 94.7% at 50% cost.
Codified Context : This study demonstrates a three-tier memory system across 283 sessions enabling agents to build a 108,000-line distributed system using 19 specialist agents and 34 knowledge documents.
Discovering Multi-Agent Learning Algorithms with LLMs : DeepMind’s AlphaEvolve evolves novel algorithms like VAD-CFR and SHOR-PSRO, outperforming baselines in imperfect-information games by autonomously designing learning logic beyond human intuition.
Global AI ethics centered on the New Delhi Declaration, where 91 nations committed to inclusive, trustworthy innovation for the Global South. Simultaneously, the European Commission proposed a “Simplification Package” for the EU AI Act, adjusting high-risk timelines. The UN Rights Chief urged human rights-based guardrails globally, while MWC 2026 launched with a focus on inclusive intelligence.
Thank you for tuning in to this week’s edition of Superintelligence Newsletter! Stay connected for more groundbreaking insights and updates on the latest in AI and superintelligence.
For more in-depth articles and expert perspectives, visit our website | Have feedback? Provide feedback.
To explore Superintelligence Media : Explore Here
Stay curious, stay informed, and keep pushing the boundaries of what’s possible!
Until Next Time!
Superintelligence Team.
The framing of "algorithmic power" is exactly right, and the juxtaposition of the New Delhi Declaration (91 nations committing to inclusive AI) with the EU simplifying AI Act timelines tells the whole story: governance is already struggling to keep pace with deployment.
What strikes me is the gap between Jamie Dimon's "prepare early" advice and what preparation actually looks like. Most workforce transition programs assume linear disruption, but multi-agent systems (like the DeepMind paper you covered) suggest the disruption curve is exponential. By the time companies finish retraining programs designed for today's AI, the AI will have moved two generations ahead.
The real question isn't whether autonomous AI should be deployed in warfare. It's whether any institution, military or civilian, has the feedback loops fast enough to govern systems that improve themselves between policy cycles.