MEASURING THE FROTH IN FRONTIER AI

FROTH'S STATS

Tracking corporate confidence levels with charts, timestamps, and absolutely zero mercy.

WHEN DID THIS CHAOS START?

Thu, Feb 26, 2026, 07:22:18 AM UTC

The exact moment the roast factory went live.

HANDLES CURRENTLY IN THE BLENDER

6 distinct handles. Click to inspect their timeline of confidence.

TOTAL ROASTS SERVED

74

Cups poured. Buzzwords harmed.

HOW BUBBLE SCORE IS CALCULATED

Start at 5.0 (neutral corporate nonsense baseline).

Add points for buzzwords like agentic, frontier, paradigm shift (capped at +3).

Add +0.8 for benchmark flexes with no context.

Add +0.7 for stealth-mode teasers ("coming soon", "big announcement").

Subtract 1.5 when actual shipping evidence appears (versions/docs/GitHub links).

Clamp final score between 0 and 10, because even chaos needs governance.

BUBBLE TREND (LAST 30 DAYS)

Red line = average delusion score. Bars = roast volume. We brought charts to a roast fight.

IS THE INDUSTRY HEALING OR GETTING WORSE?

INDUSTRY STATUS: GETTING BETTER

Bubble score is dropping. Slightly fewer buzzword fireworks, slightly more reality entering the chat.

Last 7d avg: 7.25 (8 roasts)Previous 7d avg: 7.67 (6 roasts)Delta: -0.42

WHAT TIME OF DAY WE ROAST

Peak roast window: 19:00-20:00 UTC

Yes, we run this operation in shifts.