🎧 Listen to this article

Narrated by Talon · The Noble House

The Supreme Court ruled 6-3 on Friday, February 20, that Trump exceeded IEEPA authority to impose broad tariffs. The market celebrated for 36 hours. Then Trump announced a 10% across-the-board replacement tariff. By Sunday morning it was 15%. The court's decision did not end the trade war. It forced the administration to find a new legal hook, and the administration found one before Monday open. *(Supreme Court, Learning Resources Inc. v. Trump, February 20, 2026)*

What the ruling actually resolved is narrower than the coverage suggests. It killed the specific IEEPA vehicle. It left the underlying political will intact. For businesses, that is the more important finding: tariff risk did not go away on February 20. It changed legal vehicles and repriced slightly downward.

Three signals define what that means for builders, operators, and anyone reading the economic weather.

Five Signals

1. SCOTUS rebuffed Trump on tariffs. He raised them anyway.

Editorial illustration

On February 20, the Supreme Court ruled 6-3 that Trump exceeded IEEPA authority on broad tariffs. *(Skadden Arps, February 2026)* Two days later, Trump announced replacement tariffs under Section 122 of the Trade Act of 1974, a provision never previously invoked for this purpose, starting at 10% and raised to the statutory maximum of 15% before markets opened Monday.

The mechanism: IEEPA was the vehicle for the broadest tariff authority. Killing it constrains scope and speed, not intent. Section 122 replacement authority caps at different levels and requires a different evidentiary record. The administration is betting that the supply chain reshuffling already triggered is durable enough to survive the legal downgrade.

For businesses: 15% tariffs as a floor, not a ceiling, is now the correct planning posture. The legal basis changed. The trajectory did not.

2. The creator economy is cracking, and it is not AI's fault yet.

Analysis

The global creator economy is valued at over $250 billion. The people at the center of it earn less per hour than three years ago. Fewer than 5% of full-time creators earn more than $100,000 annually. *(Influencer Marketing Hub, Creator Economy Report 2025)*

The structural reason: short-form video generates high engagement but pays fractions of a cent per view versus $5 to $15 CPM for long-form. Platforms capture the engagement. Creators absorb the production cost. TikTok pays approximately $0.02 to $0.04 per 1,000 views. YouTube long-form pays $3 to $10 per 1,000 views. The platform architecture routes attention toward the format that extracts the most engagement at the lowest payout.

The creator economy is not dying. The middle is dying: creators who built on platform-dependent ad revenue without owning their audience. When AI tools lower production costs toward zero, quality, specificity, and point of view become the only defensible positions. This is the same story as SaaS, one vertical later. The builders who diversified off platform revenue before they needed to will look prescient. The ones who did not are learning the lesson now.

3. Salesforce built the report card for AI agents.

Perspective

Salesforce AI Research released MCPEval this week, an open-source framework using the Model Context Protocol architecture to evaluate AI agent performance on tools. Shelby Heinecke, senior AI research manager at Salesforce: "We've gotten to the point where a lot of us have figured out how to deploy agents. We now need to figure out how to evaluate them properly." *(GitHub: SalesforceAIResearch/MCPEval, February 2026)*

MCPEval generates tasks dynamically, collects interaction trajectories, and builds benchmarks automatically. It captures how an agent reasons through tool use, not just whether it succeeds. The output is evaluation data and fine-tuning data simultaneously, which means organizations using it improve their agents while measuring them.

The bottleneck in agentic AI is not capability. It is trust. Operators cannot deploy agents at scale without confidence in the consistency of their outputs. Evaluation infrastructure is trust infrastructure. MCPEval builds on the same MCP protocol the agent ecosystem is converging on, which means its adoption curve will follow the broader agent adoption curve. As agents scale, the tools to measure them need to be ready. This is one of the first serious attempts to make that infrastructure available to the full community, not just the largest enterprise teams.

Compass forecast: February 23, 2026

Double-consolidation day. Foundation energy amplified. Three independent catalyst patterns active simultaneously: a rare triple convergence correlated with deal flow, coalition formation, and creative output.

Best windows (PST): 12:00 AM to 2:00 AM (creative breakthrough, SE) and 4:00 AM to 6:00 AM (deep-focus execution and strategic positioning, SW). Low-signal periods cluster at 2 AM, 6 AM, 10 AM, 12 PM PST. Avoid major launches or high-stakes public decisions during these windows. All four compass domains, wealth, health, growth, positioning, are aligned with foundation work today.

TNH call: Foundation day. Agreements, creative launches, infrastructure. The court drew a line. Figure out what you are building, not what you are fighting over.

Tomorrow watch

Tariff replacement timeline: speed of Commerce Department filings indicates how politically durable the new Section 122 authority is. If challenges are filed within days, the administration's confidence in the replacement vehicle is lower than the public posture suggests.

MCPEval adoption: GitHub stars, forks, and whether major agent framework teams issue compatibility statements within the week. Early adoption patterns for evaluation frameworks predict the tools that become de facto standards.

Creator economy earnings: major platforms reporting this week. Watch short-form versus long-form monetization rate disclosures. Any platform that reduces its stated CPM will accelerate the creator exodus to owned channels.

The Supreme Court drew a line. The executive branch stepped over it before the ink dried. Platforms drew monetization rules; short-form video found the workaround; mid-tier creators are absorbing the cost. Agents are being deployed before the industry can measure whether they work. Same pattern, different verticals. Build something that lasts.

Sources