ECONOMY - AI Dynamics

Databricks Unity Gateway Governs Usage Across AI Models and Agents

By

–

28 August 2026 21h00

More AI tools usually means more questions: who’s using what, what’s it costing, and is it actually working well?@matei_zaharia and @kuhlenhuth show how Unity Gateway helps teams see and control AI usage across models, agents, harnesses, and workflows.

They also cover how… pic.twitter.com/Q2cYnIcaks
— Databricks (@databricks) 28 août 2026

More AI tools usually means more questions: who’s using what, what’s it costing, and is it actually working well? @matei_zaharia and @kuhlenhuth show how Unity Gateway helps teams see and control AI usage across models, agents, harnesses, and workflows. They also cover how

→ View original post on X — @databricks

28 August 2026

Unify Reports Cutting AI Agent Costs by 95 Percent

By

@langchain

–

28 August 2026 20h01

Watch or listen to the latest Max Agency on your favorite podcasting platform. Apple: https://
podcasts.apple.com/us/podcast/how
-unify-cut-its-ai-agent-costs-95-in-two-weeks/id1891551672?i=1000783151404
… Spotify: https://
open.spotify.com/episode/6kWQou
c2QmiHGk0vdiZEtd?si=ba6e241ca4a24faf
… YouTube: https://
youtu.be/6898VdRtKDE

→ View original post on X — @langchain

28 August 2026

Unify Reports Cutting AI Agent Compute Costs by 90–95 Percent

By

@langchain

–

28 August 2026 20h01

Two weeks before launch, @unifygtm's agent was burning so much compute that one message could instantly eat through a customer budget.

Co-Founder & CTO @HeggieConnor on how they cut costs by 90-95% pic.twitter.com/rrYIBFup1q
— LangChain (@LangChain) 28 août 2026

Two weeks before launch, @unifygtm
's agent was burning so much compute that one message could instantly eat through a customer budget. Co-Founder & CTO @HeggieConnor on how they cut costs by 90-95%

→ View original post on X — @langchain

28 August 2026

Polsia Agents Manage Email, Stripe, Code and Ad Accounts

By

@testingcatalog

–

28 August 2026 19h54

Users have full access to the same inbox and negotiation tools, with agents handling business emails, Stripe management, codebase updates, and connected ad accounts on a daily cycle. Polsia reports over 1,000 companies running on the platform and an annual run rate approaching

→ View original post on X — @testingcatalog

28 August 2026

Meta’s Muse Image Launch Targets Production Image Generation Costs

By

@alexandr_wang

–

28 August 2026 19h28

1/ muse image is live on the meta model api $0.01/image – one of the best price-to-quality ratios for production volumes

→ View original post on X — @alexandr_wang

28 August 2026

CommerceAgentBench Includes Noisy Procurement Data and Payment Redirection Attempts

By

@testingcatalog

–

28 August 2026 18h36

Tasks pulled from a real operational mess > 300 noisy procurement emails with a payment redirection attempt buried inside. > A browser listing form that rejects bad input. > A twelve-day supplier itinerary seeded with ten classes of human error. Fourteen mock services ship

→ View original post on X — @testingcatalog

28 August 2026

CommerceAgentBench Scores Agents by Changes Left in Mock Commerce Systems

By

@testingcatalog

–

28 August 2026 18h36

Accio has open-sourced CommerceAgentBench, a 107-task benchmark spanning procurement, product listing, operations, order fulfillment, and after-sales. The benchmark ranks scores based on the changes an agent leaves behind in a mock commerce stack, such as labels applied, drafts

→ View original post on X — @testingcatalog

28 August 2026

AI Growth Raises Questions About Concentration of Wealth

By

@bamitav

–

28 August 2026 15h14

AI is no longer just a technology sector. It is becoming an economic force. But the bigger question is: will AI-driven growth spread across economies and businesses, or concentrate wealth among a handful of companies and countries? That may define the next decade.

→ View original post on X — @bamitav

28 August 2026

GLM-5.3 Local FP8 Inference Can Require Multiple H100 GPUs

By

@kimmonismus

–

28 August 2026 13h12

you just need 10–12× H100 for FP8, or $400.000–$650.000 to run GLM-5.3 locally.

→ View original post on X — @kimmonismus

28 August 2026

CommerceAgentBench Tests Agents on 107 Open E-Commerce Tasks

By

@kimmonismus

–

28 August 2026 10h15

Most AI benchmarks test whether a model can give the right answer. CommerceAgentBench asks whether an agent can actually finish the work. Accio has open-sourced 107 e-commerce tasks across procurement, product listings, operations, fulfillment and after-sales. Agents work across

→ View original post on X — @kimmonismus

28 August 2026