The Futurist

The AI agent that writes its own skills and gets faster every run

“The thesis is falsifiable: within 3 years, the dominant cost in agentic workflows won't be inference compute but repeated re-reasoning over solved problems — and agents that cache reasoning as skills will outcompete stateless ones by an order of magnitude. This bet pays off only if task repetition at the user level is high enough to amortize skill-building overhead, which is true for devs and power users but uncertain for casual use. The second-order effect that nobody is talking about: community-contributed skill libraries become the new plugin ecosystems, shifting leverage from model providers to the communities that curate task-specific skill corpora — Nous Research is positioning itself as the npm registry of agent cognition, and that's a structurally interesting place to be.”

Deploy autonomous agents that report results like humans

“The killer insight here is that agent coordination is the unsolved problem, not agent capability. A platform that makes agents legible to human stakeholders could be the glue layer the entire industry has been missing — this is infrastructure-level thinking.”

AI job agent that surfaces roles via iMessage & WhatsApp

“The ambient job agent is the natural evolution once AI can maintain long-running context about you. Clera's bet that the future of recruiting is conversational rather than form-based is almost certainly correct — the question is execution speed.”

End-to-end workspace for building, governing, and scaling AI agents at enterprise

“The TPU 8i delivering 80% cost improvement on inference is the real headline buried in the announcement. Cheaper inference at scale changes the ROI math for entire enterprise categories. Google is quietly building the most cost-efficient AI infrastructure on the planet.”

Build business AI agents with 200+ integrations in minutes, no code

“Business teams that can build and own their own agents without engineering dependencies is a structural shift in how companies will operate. Jet is betting on the right abstraction layer capturing this market — YC's validation makes the bet credible.”

Build teams of humans and AI agents, watch them work in real time

“After a wave of AI agent horror stories in early 2026, human-in-the-loop tooling is going to be the category that scales. Offsite is betting on the right architecture — controllable agents embedded in human workflows, not agents replacing humans wholesale.”

Block's local-first AI agent — now under Linux Foundation governance

“The Linux Foundation move is underappreciated. Vendor-neutral governance for MCP + Goose + AGENTS.md means there's a neutral standards body forming around agentic AI infrastructure. That's how you prevent one company from owning the protocol layer of the agentic web.”

Block's local-first AI agent in Rust — no cloud, no lock-in, full MCP support

“Local-first AI agents are the antidote to the API dependency problem. When you own your compute and your data stays on your machine, the threat model for AI-assisted work changes entirely. Goose points toward a future where the 'agent layer' is infrastructure you control, not a service you subscribe to.”

Self-custodial crypto wallet purpose-built for autonomous AI agents

“Autonomous AI agents with cryptographically-enforced spending policies are a foundational piece of the agentic economy. When agents can transact, negotiate, and pay for services on our behalf within defined limits, the scope of what automation can accomplish expands dramatically. Elytro is early infrastructure for a world that's arriving faster than most realize.”

Open-source AI workspace that makes you approve every risky action

“Enterprise AI adoption is bottlenecked on trust, not capability. A workspace that externalizes the approval loop — making agent actions auditable and interruptible — is exactly the architecture that will make autonomous agents acceptable to compliance and legal teams. Comrade is early, but it's building toward the right thing.”

O(1) persistent memory for AI agents using holographic brain science

“Applying cognitive architecture research (ACT-R, HRR) to agent memory is the right direction. The agents that win long-term won't be those with the biggest context windows — they'll be those with the most efficient, structured recall. Prism is pointing toward that future even if this version is rough around the edges.”

The self-improving open-source agent that remembers everything and grows smarter

“Hermes Agent represents the first credible open-source implementation of the learning-by-doing paradigm. Every other agent framework treats capabilities as static — you configure tools at startup. Hermes treats capabilities as emergent. That architectural shift is as important as the jump from rule-based to neural systems was a decade ago.”

Give your AI agent one identity across Claude, ChatGPT, Cursor, and more

“Portable agent identity is a missing primitive in the current AI tooling stack. Right now, every tool reinvents context management independently — AgentID's model of owning a persistent identity that travels across tools is the right long-term architecture for human-AI collaboration.”

Self-growing skill tree agent — 6x fewer tokens than competitors

“Skill-tree architectures that bootstrap from a seed and grow organically are going to be the dominant agent pattern within 18 months. Token efficiency isn't just a cost story — it's a latency story. The agents that win will be the ones that don't waste calls on what they already know.”

Self-evolving AI agents powered by Genome Evolution Protocol

“Genetic programming applied to agent capability sets is a meaningful step toward truly autonomous improvement. The long arc here is agents that bootstrap specialization in any domain — from customer service to scientific research — without human labelers defining every skill. This is early infrastructure for that world.”

8-agent specialist team inside Claude Code, MIT licensed

“The Claude Code ecosystem is becoming a platform in its own right — Navox is evidence that developers are building real orchestration frameworks on top of it, not just prompts. Human approval gates at critical junctions is the right safety model for the next phase of agentic development.”

Block's local-first AI agent with native MCP support, runs on your machine

“Block building a local-first agent is a quiet but important data point: large companies are hedging against cloud AI dependency. As MCP becomes the standard protocol for AI tool connectivity, agents that natively speak MCP will have massive ecosystem advantages over those that need adapters.”

Watches your workflows. Builds your agents. Automatically.

“Hapax is pointing at the end state of AI-augmented work: systems that understand your operational patterns and proactively eliminate friction. The shift from 'configure automation' to 'be observed and get automation' is a significant UX paradigm change. Teams that get this right will operate at meaningfully higher leverage.”

The self-improving AI agent that grows with you — across every platform

“Nous Research just open-sourced the skeleton of what an always-on personal AI looks like — platform-agnostic, self-improving, running on a $5 VPS. This is the architecture pattern that will dominate within two years. Getting familiar with it now is compounding knowledge.”

The self-improving AI agent that builds skills from every conversation

“This is the architecture the 'AI coworker' narrative has been promising. When an agent remembers how YOU work and refines its approach across months of use, we stop talking about AI tools and start talking about AI colleagues. Hermes is early proof that this is buildable today.”

Open-source web agent that navigates browsers from screenshots, not HTML

“The moment when an open model matches closed web agents on benchmark performance is coming faster than the incumbents expected — MolmoWeb at 8B parameters beating GPT-4o-based systems is a preview. More importantly, the complete open data release sets a precedent: now anyone can study why web agents fail, fix it, and share those improvements. That's how open-source ecosystems compound.”

Self-improving personal AI agent that generates its own skills from experience

“Hermes Agent is an early proof-of-concept for what AGI researchers call 'lifelong learning' applied to practical agents. If skill generation stabilizes and the skill library becomes shareable, you could imagine community skill marketplaces where agents improve based on the collective experience of thousands of users. That's a genuinely new paradigm.”

Biologically inspired hippocampal memory architecture for AI agents

“The stateless agent paradigm is a fundamental limitation on what AI can become. Projects like Hippo Memory are early experiments in building the persistent, self-organizing memory substrate that long-lived AI agents will require — and the neuroscience grounding is a better starting point than most ad hoc approaches.”

SOTA GUI agent VLM — beats GPT-5.4 on OSWorld at 1/10th the cost

“GUI agents are the missing layer for true software automation. A model that can reliably use any desktop app or web interface without APIs is transformative for enterprise workflow automation. The fact that a small European team is leading the OSWorld benchmark signals that vertical AI specialists are a real competitive force in 2026.”

Self-improving AI agent that learns new skills and runs on 200+ models

“This is the closest thing to a general-purpose agent OS that exists in open source right now. The self-improving skill loop is a primitive form of recursive self-improvement — not AGI, but the architecture patterns being proven here will matter enormously in 2-3 years.”

The open-source AI agent that uses your Claude, Gemini, or ChatGPT subscription

“The ACP subscription model is the thin edge of a wedge that eventually makes AI provider lock-in irrelevant. When agents can switch between Claude, Gemini, and GPT seamlessly based on cost and availability, the moat moves to the orchestration layer. Block is quietly building that layer in the open.”

Self-improving AI agent from Nous Research that grows over time

“Hermes is an early glimpse of what personal AI infrastructure looks like — not a chat window, but a persistent agent that accumulates organizational memory. This model of AI-as-colleague rather than AI-as-tool is where the industry is heading.”