Forecasts and trend signals from the gentic.news knowledge graph.
Every prediction was written by the brain after reading the news, scored 0–100 for confidence, and resolves automatically against future evidence. Sort by confidence to see strongest signals; switch to Resolved to grade past calls.
AI-generated predictions backed by knowledge graph analysis of 89+ news sources. Each prediction cites specific entities, relationships, and trend signals — then gets automatically verified against real outcomes.
Each one is a falsifiable claim with a deadline and a confidence score. We watch the news, log the outcome, and report calibration honestly — including when we’re wrong.
X = stated confidence. Y = how often we were right. The diagonal is perfect calibration.
Within the next quarter, Google Cloud will make at least one agentic coding or workflow tier bill separately from core Gemini usage, either through distinct metering, a dedicated SKU, or a usage policy that clearly decouples agent actions from raw model tokens. The tell will be that Google starts pricing the workflow layer, not just the model layer.
Within the next quarter, GitHub will publicly ship a first-party MCP gateway or policy layer for Copilot-style workflows. The feature will be positioned around connector approval, tool allowlists, and auditability rather than raw model quality.
Within the next quarter, Google will ship a Chrome-integrated Gemma 4 experience that can complete multi-step browser tasks locally or with minimal cloud roundtrips. The key tell will be that Google positions Gemma not just as a model, but as the default execution layer for consumer browsing workflows.
Google Cloud will expose at least one TPU pricing path that is explicitly cheaper for inference-heavy or agentic workloads than its current general-purpose model hosting. The practical effect will be to make TPU economics legible to developers in the same way GPU spot pricing already is, and to pressure model providers that rely on Google infrastructure to repackage their own margins.
Within the next quarter, Microsoft will expose at least one Copilot agent capability under a separate enterprise billing or metering path instead of bundling it into standard Microsoft 365 pricing. The tell will be a distinct SKU, usage meter, or admin-controlled quota for agentic workflows rather than a generic Copilot add-on.
Within the next quarter, Google will make at least one TPU pricing or billing path explicitly distinct for agentic or workflow-heavy inference, not just generic model usage. The practical signal will be a separate SKU, calculator, or documented rate card that makes long-running tool-using workloads cheaper or easier to meter than standard Gemini calls.
Within the next quarter, Google will introduce a materially cheaper Gemini tier or usage policy aimed specifically at coding and agentic workflows. The move will be framed as developer-friendly pricing, but the real target will be Claude Code and OpenAI’s coding stack.
Within the next quarter, Google Cloud will expose at least one materially cheaper pricing path for agentic or workflow-heavy usage than its standard Gemini API tiers. The move will be framed less as a model launch and more as a billing distinction for long-running, tool-using workloads that need predictable cost envelopes.
Your vote stays in your browser. We compare crowd intuition against the lab’s calibrated forecast.
Within the next quarter, Google Cloud will make at least one agentic coding or workflow tier bill separately from core Gemini usage, either through distinct metering, a dedicated SKU, or a usage policy that clearly decouples agent actions from raw model tokens. The tell will be that Google starts pricing the workflow layer, not just the model layer.
Top movers from 7-day mention velocity.
Auto-verified (confidence=85%, corroboration=72%, threshold=75%, web_search=yes): The prediction that Anthropic will release native GitHub integration for Claude Agent is substantively correct. Anthropic's official platform documentation ([W6]) explicitly describes connecting agents to GitHub for cloning, reading, and creating pull requests. The official 'claude-code-action' GitHub repository ([W7]) provides PR analysis, code implementation, and issue access. While no formal blog post was found, the verification criteria allow for 'developer documentation,' which these official sources fulfill. The launch of Claude Managed Agents ([W1]) provides the service infrastructure. The prediction's core claims—repository access, PR automation, and codebase analysis—are all confirmed by primary Anthropic sources. [Evidence FOR (4): [W6] Anthropic's platform documentation at platform.claude.com shows a dedicated page for 'Accessing GitHub' under Managed Agents, confirming that agents can 'Connect your agent to GitHub repositories for cloning, reading, and creating pull requests' and 'mount a GitHub repository to your session container and connect to the GitHub MCP for making pull re...'; [W7] Anthropic's official GitHub repository features 'claude-code-action', an interactive code assistant that 'Analyzes PR changes and suggests improvements', 'Can implement code changes and create commits/PRs', and 'Accesses GitHub issues, PRs, and code context', directly fulfilling the PR automation and codebase analysis criteria.; [W1] SiliconAngle reports on April 8, 2026 that 'Anthropic launches Claude Managed Agents to speed up AI agent development', a cloud service that likely underpins the GitHub integration. | Evidence AGAINST (3): No evidence found of an official Anthropic blog post specifically announcing a native GitHub integration for Claude Agent, though the verification criteria allow for 'developer documentation' which [W6] fulfills.; [W4] The Verge reports on a Claude Code source code leak showing unreleased features, but none of the leaked features described include a native GitHub integration; this absence is weak evidence against.]
Auto-expired: past deadline, inconclusive (confidence=0%, corroboration=0%, web_search=yes)
Auto-verified (confidence=85%, corroboration=75%, threshold=85%): The key evidence [DB-11] directly confirms the core prediction: Anthropic removed Claude Code from the $20/month Pro plan and moved it to $100+ tiers, making it materially less standalone and pushing heavy users toward higher-tier plans. This is a visible pricing/bundling change that shifts the economic center of gravity away from a simple developer tool. While some evidence shows continued standalone use, the pricing change is a concrete manifestation of the predicted tightening. The corroboration score is high because [DB-11] is a credible source (Towards AI, aggregated across multiple feeds) and directly matches the verification criteria. [Evidence FOR (4): [DB-11] Anthropic Removes Claude Code from $20 Plan, Signals AI Pricing Shift — confirms Claude Code was removed from the $20/month Pro plan and moved to $100+ tiers, directly supporting the prediction of tightened bundling and pricing changes.; [DB-11] The same article notes this reflects high operational costs and signals a pricing shift, aligning with the prediction that economic center of gravity shifts away from a simple developer tool.; [DB-1] Anthropic published a post-mortem on Claude Code quality issues, indicating active management and potential tightening of access/usage rules. | Evidence AGAINST (4): [DB-0] Cua open-sourced a driver that allows Claude Code to drive macOS apps, suggesting continued standalone utility and ecosystem growth.; [DB-3] AgentBox SDK allows running Claude Code in any sandbox, indicating the tool remains flexible and standalone for developers.]
Auto-verified (confidence=90%, corroboration=85%, threshold=75%, web_search=yes): The prediction is verified by OpenAI's official blog post (WEB-6) announcing 'Powering Product Discovery in ChatGPT' with the Agentic Commerce Protocol for product discovery, comparison, and merchant integration, meeting the verification criteria of an official announcement. This is corroborated by a third-party article (WEB-7) detailing Stripe's integration for AI shopping in ChatGPT. The evidence confirms the substance of the prediction—commerce-specific capabilities—before the May 31, 2026 deadline, with authoritative sources providing clear confirmation. [Evidence FOR (4): [WEB-6] OpenAI blog post titled 'Powering Product Discovery in ChatGPT' announces ChatGPT introduces richer, visually immersive shopping powered by the Agentic Commerce Protocol, enabling product discovery, side-by-side comparisons, and merchant integration.; [WEB-7] Article 'Stripe's Agentic Commerce Suite Powers AI Shopping in ChatGPT ...' confirms Stripe's Agentic Commerce Suite lets brands sell through ChatGPT and Copilot via one integration, indicating a commerce-specific feature with checkout integration.; [W1] Zendrop launches a Model Context Protocol (MCP) server that gives AI assistants like ChatGPT the ability to run a store, supporting the idea of commerce integration for ChatGPT. | Evidence AGAINST (3): [DB-0] to [DB-24] No database articles mention an OpenAI commerce-specific API or agent feature; all are about unrelated topics like health advice, user growth, or competitor releases.; [DB-9] OpenAI shifts ChatGPT ads to CPC, targeting ad revenue, but this is about advertising, not a commerce-specific API for product search/checkout.]
Auto-verified (confidence=85%, corroboration=41%, threshold=85%): The prediction's core claim is the launch of Qwen 4.0 at the Alibaba Cloud Summit. Multiple database news items from April 2026 ([DB-1], [DB-2], [DB-4]) explicitly refer to Qwen 3.6 as the latest released model, with one calling Qwen 3.6 Plus the current 'frontier model.' This directly contradicts the existence of a launched Qwen 4.0. While evidence shows Alibaba is active in AI and the Qwen series, the specific predicted entity (Qwen 4.0) has not materialized. The deadline for the 'typically June' summit has not passed, but the evidence shows a different, contradictory reality (Qwen 3.6 is the current version), moving the judgment from 'inconclusive' to 'incorrect.' [Evidence FOR (4): [DB-11] Alibaba's Qwen Hits 1B Downloads, Captures 50% of Open-Source Market (April 10, 2026). This shows the Qwen family is active and successful, providing context for a future major release.; [DB-1] Alibaba Makes Qwen 3.6 Plus API-Only, Shifts Frontier Model to Paid Access (April 19, 2026). This indicates a strategic shift towards monetizing advanced models, aligning with a potential premium Qwen 4.0 launch.; [DB-2] Qwen 3.6 Released: Free, Open-Weights Model for Local AI Coding (April 17, 2026). This confirms ongoing development and release of the Qwen series, with version 3.6 being the latest announced model. | Evidence AGAINST (3): [DB-2] Qwen 3.6 Released: Free, Open-Weights Model for Local AI Coding (April 17, 2026). This directly contradicts the prediction, as the latest announced model is Qwen 3.6, not Qwen 4.0.; [DB-1] Alibaba Makes Qwen 3.6 Plus API-Only... (April 19, 2026). This discusses Qwen 3.6 Plus as the current 'frontier model,' with no mention of Qwen 4.0.]
Auto-verified (confidence=75%, corroboration=65%, threshold=75%, web_search=yes): The prediction specified that Claude Code would publicly ship a mode for autonomous pull-request workflows (fixing CI, responding to reviews, opening follow-up PRs) within a month, shifting it to an 'always-on repo operator.' Evidence shows Anthropic launched 'Routines' for Claude Code in mid-April 2026, which is an automation feature described as moving beyond interactive assistance. However, the verification criteria require a specific background/autonomous PR workflow for the narrow tasks listed, and the articles about 'Routines' do not explicitly confirm it handles CI fixes, review responses, or follow-up PRs end-to-end. The substance is partially met—Claude Code gained automated capabilities—but the specific predicted workflow is not verified. [Evidence FOR (5): [W0] VentureBeat article (April 14, 2026) reports Anthropic launched 'Routines' in research preview with the Claude Code desktop app redesign, describing it as a shift toward automation.; [W1] SiliconANGLE article (April 14, 2026) confirms the launch of 'Routines' in Claude Code, stating it allows automation of tasks without relying on autonomous AI agents.; [W2] Thurrott article reports a redesigned Claude desktop app supporting parallel agents for running more Code tasks simultaneously. | Evidence AGAINST (2): [DB-0], [DB-1], [DB-2], [DB-3], [DB-5], [DB-6], [DB-7], [DB-8], [DB-9], [DB-10], [DB-11], [DB-13], [DB-14], [DB-15], [DB-16], [DB-17], [DB-18] — None of these database articles mention the specific predicted feature (autonomous PR workflow for fixing CI, responding to reviews, or opening follow-up PRs).; [W3], [W4], [W5], [W6], [W7] — Web search results discuss leaks, Mac control, or general news but do not confirm the specific narrow pull-request workflow capability.]
Top 30 anonymous voters · ranked by accuracy on resolved predictions
Predictions are generated by analyzing trend signals across 42+ AI news sources, enriched with knowledge graph relationships between entities (companies, people, technologies). Each prediction includes a confidence score and target date.
Accuracy = (correct + partial × 0.5) ÷ total evaluated. All resolved predictions count — including expired ones (treated as failures). Sample size is shown next to the accuracy figure.
Past-deadline predictions are verified via 3-layer evidence: entity-linked articles, keyword search, and web search. An AI judge evaluates evidence for and against, requiring high confidence thresholds before resolving.
Within the next quarter, Google Cloud will expose at least one materially cheaper pricing path for agentic or workflow-heavy usage than its standard Gemini API tiers. The move will be framed less as a model launch and more as a billing distinction for long-running, tool-using workloads that need predictable cost envelopes.
Within the next quarter, at least two MCP-focused security products will move beyond connector discovery and ship policy enforcement features such as allowlists, tool-level approvals, or session logging. The market will stop selling MCP as an integration problem and start selling it as an enterprise control plane problem.
Within the next quarter, at least two developer-facing AI products in Asia will reprice or repackage around DeepSeek-style cost efficiency, not frontier capability. The interesting part is not that DeepSeek gets adopted, but that it forces local wrappers and middleware to stop selling 'best model access' and start selling 'lowest-cost acceptable automation.'
Within the next quarter, Google will ship a Chrome-integrated Gemma 4 experience that can complete multi-step browser tasks locally or with minimal cloud roundtrips. The key tell will be that Google positions Gemma not just as a model, but as the default execution layer for consumer browsing workflows.
Google Cloud will expose at least one TPU pricing path that is explicitly cheaper for inference-heavy or agentic workloads than its current general-purpose model hosting. The practical effect will be to make TPU economics legible to developers in the same way GPU spot pricing already is, and to pressure model providers that rely on Google infrastructure to repackage their own margins.
Within the next quarter, Microsoft will expose at least one Copilot agent capability under a separate enterprise billing or metering path instead of bundling it into standard Microsoft 365 pricing. The tell will be a distinct SKU, usage meter, or admin-controlled quota for agentic workflows rather than a generic Copilot add-on.
Within the next quarter, Cursor will ship a first-party MCP policy or connector-management layer aimed at enterprise teams. The tell will be admin controls for allowed tools, connector approval, or auditability rather than another model-quality feature.
Within the next quarter, Google will make at least one TPU pricing or billing path explicitly distinct for agentic or workflow-heavy inference, not just generic model usage. The practical signal will be a separate SKU, calculator, or documented rate card that makes long-running tool-using workloads cheaper or easier to meter than standard Gemini calls.
Within the next quarter, OpenAI will reduce effective pricing or expand usage limits for at least one coding-relevant API tier, but it will not do so through a broad ChatGPT discount. The move will be narrowly aimed at developer retention, not consumer growth, and will look more like a tactical API response than a product reset.
Within the next quarter, Google Cloud will make at least one agentic coding or workflow tier bill separately from core Gemini usage, either through distinct metering, a dedicated SKU, or a usage policy that clearly decouples agent actions from raw model tokens. The tell will be that Google starts pricing the workflow layer, not just the model layer.
Within the next month, OpenAI will make Codex materially more distinct from ChatGPT in pricing or packaging, with a separate developer-facing billing surface or usage tier. The practical result will be that coding-heavy customers stop being treated as generic ChatGPT users and start being sold a dedicated workflow product.
Within the next quarter, at least two security vendors will launch MCP-specific enterprise controls such as connector approval, tool logging, or policy enforcement. The market will form around the uncomfortable fact that the same protocol making agents useful also makes them governable only if someone owns the control plane.
Within the next quarter, Google will introduce a materially cheaper Gemini tier or usage policy aimed specifically at coding and agentic workflows. The move will be framed as developer-friendly pricing, but the real target will be Claude Code and OpenAI’s coding stack.
Within the next month, Anthropic will make Claude Code materially more distinct from Claude AI in pricing or billing, with a separate seat, usage, or enterprise packaging layer. The change will not just be cosmetic: heavy coding users will be pushed into a different commercial bucket than general Claude users.
Within the next quarter, GitHub will publicly ship a first-party MCP gateway or policy layer for Copilot-style workflows. The feature will be positioned around connector approval, tool allowlists, and auditability rather than raw model quality.
Within the next quarter, at least two AI middleware or wrapper startups will publicly reprice, repackage, or downshift margins after DeepSeek’s next open-weights releases make inference economics visibly worse for them. The hidden dynamic is that cheap frontier-quality open models don’t just pressure closed labs — they compress the value of orchestration layers that were charging for access to model capability rather than workflow depth.
Within the next quarter, DeepSeek will ship or describe a next-step model pipeline that relies primarily on synthetic data generated by its own prior model family. The interesting part is not just synthetic data use, but the first clearly productionized self-improvement loop from a major open-weight challenger.
Qwen 3.6 Plus is the strongest surge in the graph, and its cascade hits Alibaba, Qwen2-VL-2B, and Qwen3-Coder-Next. That pattern implies Alibaba is not treating it as a standalone model win; it is being used to seed adjacent coding and multimodal products, likely with a public developer release or benchmark push.
Join readers from Google, Anthropic, and NVIDIA. Every week: the 10 most important AI developments, verified predictions, and what they mean for your work. Free forever. Customize what you get →