Nik

Farther of 2. AI Builder (TACIT app). EIR in enterprise, bringing HRtech AI expertise to VC/PE. My own opinions & random quotes (#quoteserious). Angel investor.

[Nik @mycomradio]

The Agentic AI Platform War: Who Controls Enterprise Memory, Context, and Action?
WindowsNews published a definitive analysis: the battle has shifted from "who has the best model" to who owns the agentic client — the chokepoint controlling enterprise memory, context, and execution. Microsoft, Snowflake, Databricks, Google, OpenAI, Anthropic, Salesforce, SAP all converging. Governance is now the primary differentiator, not model capability.
→ https://t.co/xz2CIqZvwz

[Nik @mycomradio]

Tacit logo

[Nik @mycomradio]

Apple WWDC: Claude Integrated into Apple Intelligence — 2.2B Device Distribution Play
Apple rebuilt Siri with a custom 1.2T-parameter Gemini model under a ~$1B/year Google deal. But the real story: iOS 27 lets users choose ChatGPT, Gemini, or Claude as their AI backend. Claude integration gives Anthropic potential access to 2.2B Apple devices. Agentic capabilities in system apps — Passwords app auto-navigates sites and updates credentials. macOS "Golden Gate" ships with native agent capabilities.
→ https://t.co/2ubogXbZMR

[Nik @mycomradio]

@signulll Priorat is too mineral

[Nik @mycomradio]

@signulll Priorat is too mineral

[Nik @mycomradio]

Flourish — $500M at $2.5B (Brain-Inspired AI)
Led by Lux Capital and GV, with ~$100M from Jeff Bezos personally. Building brain-inspired AI architectures. The bet: current transformer architectures hit a wall on reasoning; neuromorphic approaches will power the next generation of enterprise agents.
→ https://t.co/djLqMarELC

[Nik @mycomradio]

Colrows
• Product: Semantic Execution Layer — typed, versioned semantic graph for enterprise AI
• Thesis: Every business object needs a semantic identity agents can reason over
• Differentiator: Not just storing context, but making it executable — agents can traverse the graph to find not just WHAT but WHY and HOW
• Stage: Early (no public funding announced)
• Watch because: If the "Company Brain" category consolidates, Colrows' typed semantic graph could become the schema standard

[Nik @mycomradio]

In the future everyone will come to work with their own context agent

[Allie K. Miller @alliekmiller]

Everyone's talking about "second brain" for AI.

I added a new layer to mine.

I built a context vault with 200-700 line summary docs of big areas of my life (business, 2026 goals, family, friends, a personal constitution). WAY fewer tokens than pointing Claude at a lifetime of tweets and emails.

But use case matters.

👉 "What do I think about this topic?" → you need actual notes
👉 "Customize this app for my goals" → you just need the critical distillation of context vault docs

Then I nest different docs into different tasks. Drafting a client proposal? It reviews my constitution, 2026 goals, and business docs. Booking a big travel adventure? Constitution + friends and family docs.

Your AI doesn't need your whole life for everything. It needs the RIGHT slices for the task at hand.

You can prompt on your own or grab mine here: https://t.co/JTgmc6tKlp

[Nik @mycomradio]

@paulg Status game is still very strong but the value incl network is dropping

[Nik @mycomradio]

@16vchq TACIT

[Nik @mycomradio]

@1Umairshaikh What did Claude say?

[Nik @mycomradio]

Snowflake Summit 2026 is the week "context" went from thesis to product SKU. Snowflake (Horizon Context), Atlan (Context Lakehouse), Arango (AutoGraph), RelationalAI (Rel App), and Open Semantic Exchange all shipped in the same 72 hours. The governed context layer is now the battleground. The question has flipped from "do we need context?" to "whose context layer becomes the standard?"

[Nik @mycomradio]

HackerNoon: "Context Graphs, Ontologies, and the Race to Fix Enterprise AI"

HackerNoon published a comprehensive piece mapping the ontology debate in context graphs — prescribed ontologies (Palantir-style) vs. learned ontologies that emerge from actual work patterns. The piece argues the real race isn't about which graph database wins, but which approach to capturing business meaning scales across heterogeneous enterprise environments.
https://t.co/ZFRWX1Iogm

[Nik @mycomradio]

Atlan's Context Lakehouse Hits 8 Billion Reads in 90 Days

Atlan revealed that its "Context Lakehouse" logged 8 billion reads in just 90 days — a unified context layer accessible via MCP, API, or SQL by engines like Cortex, Genie, Claude, and Codex. Their "Context Agents" compress 9-12 month metadata enrichment into ~30 days, with 50+ teams generating over 1 million AI-created descriptions in two weeks. Context Engineering Studio can move an AI agent from concept to production in under 10 minutes.
→ https://t.co/vENTpxJTln

[Nik @mycomradio]

Snowflake Summit 2026: "Context" Is Now the Product

Snowflake shipped Horizon Context — a governed context layer for AI, BI, and apps — making "context" an official product category in the enterprise data stack. Key features: Semantic Studio (define shared business logic without SQL), Semantic View Autopilot (auto-refines semantic views), and Cortex Sense (auto-transforms metadata into agent-ready ontology). BlackRock is already using Horizon Context. CEO Sridhar Ramaswamy's framing: the agentic enterprise requires four pillars — unified data, AI models, enterprise apps, and an agent control plane.
→ https://t.co/qbVTpEOztt

[Nik @mycomradio]

@emollick Has always been like this

[Nik @mycomradio]

Coupa Acquires Rossum — AI Document Processing for Supply Chain
Coupa bought Rossum to add AI document understanding to its business spend management platform. Structured extraction feeding enterprise procurement agents.

[Nik @mycomradio]

@agupta Because surveys are the most reliable source of truth.

[Nik @mycomradio]

Cognizant Opens TriZetto to AI Agents for Healthcare
TriZetto Unify now treats AI agents as first-tier API consumers, starting with Electronic Prior Authorization aligned with HL7 FHIR. Shifts healthcare automation from UI scripting to direct API interaction — a concrete example of vertical agent infrastructure going live.

[Nik @mycomradio]

Gartner: 40% of Enterprise Agent Projects Will Be Decommissioned by 2027
New Gartner prediction: enterprises without multi-tiered governance matched to agent autonomy levels will demote or kill ~40% of their agent deployments. Governance is no longer optional — it is the gating factor for enterprise agent adoption at scale.