$3M Grants & Open Benchmarks: AI Evaluations Improve 💵

Firefly boost assets while Deepfakes are testing limits

Author

Lauren Brandt
February 15, 2026

In partnership with

Better prompts. Better AI output.

AI gets smarter when your input is complete. Wispr Flow helps you think out loud and capture full context by voice, then turns that speech into a clean, structured prompt you can paste into ChatGPT, Claude, or any assistant. No more chopping up thoughts into typed paragraphs. Preserve constraints, examples, edge cases, and tone by speaking them once. The result is faster iteration, more precise outputs, and less time re-prompting. Try Wispr Flow for AI or see a 30-second demo.

Hey there,

AI is hitting everything from trucking to elections, while Firefly accelerates creative work new benchmarks are chasing agentic AI, to stocks getting unstable.

Deepfakes spread, and grant measures to keep everything in place.

Which sectors are you leveraging—and where will you draw the line?

AI TOOL SPOTLIGHT

Adobe Firefly

Adobe Firefly is an AI-powered creative space for generating images, videos, and edits from text prompts on infinite canvases, integrated with Photoshop and Premiere for generative fill, speech, and quick actions like background removal.

Best for

  • Artists and designers create realistic landscapes, portraits, or surreal visuals.

  • Teams producing marketing media, edited videos, or custom assets fast.

How to use it

  • Enter detailed prompts for images/videos like "sunrise mountain valley."

  • Edit media with a text: generative fill for empty spaces or a video trim.

  • Quick actions: removes the backgrounds, and can generate QR codes.

When not to use it
For non-creative tasks like data analysis, it’s better to stick to specialized AI tools.

Pro tip
Combine with Photoshop for workflows, refining prompts for best results.

FEATURE STORY

📈 AI's Relentless Change: Disrupts Estate, Trucking, & Beyond

CNBC reports AI's real estate, trucking, and logistics invasion, plummeting stocks as investors flee labor-heavy models for AI-based efficiency. Elon Musk predicts empty office towers as AI erases white-collar jobs, while Algorhythm Holdings' platform are scaling freight around 300-400% sans new hires. SoftBank gains $4.2B on OpenAI, as Singapore launches their own national AI council.

Key Takeaways:

  • 📉 Sector Carnage: Real-estate and trucking stocks plunge as investors rotate from AI-vulnerable, high-fee labor models.

  • 🚚 Freight Revolution: Algorhythm's tool scales freight volumes 300-400% without added staff, reeling logistics firms.​

  • 🏢 Office Doom: Musk and Shumer predict AI erasing entry-level jobs, emptying corporate and office firms, and killing leases.

  • 💰 AI Surge: Anthropic's $30B at $380B valuation, SoftBank's $4.2B OpenAI gain, and Singapore's AI council fuel momentum.

📊 Closing Agentic AI's Evaluation Gap with Open Benchmarks

Snorkel announces $3M Open Benchmarks Grants to discuss critical evaluation break in agentic AI, where development outpaces measurement beyond coding. Benchmarks must evolve across complexity, horizon, and complexity to enable high-stakes deployments. Partners like Hugging Face, Together AI, and PyTorch supports this upgrade.

Key Takeaways:

  • 🔄Eval Gap Crisis: AI development surges in agentic systems, but measurement remains lagging with high-stakes deployments.​

  • 🔬 Benchmark Evolution: Future experiments target environment complexity (domain, multi-modal, tools) & long-term autonomy.

  • 🌐 $3M Grants Launch: Snorkel has now funded open benchmark R&D with Hugging Face, Together AI, PyTorch, setting from March 1.

  • 🛡️ Safe AI Path: Advanced benchmarks like Terminal-Bench guides trustworthy agents, countering "benchmaxxing" potential risks.

🚨 AI Deepfakes from U.S. Threaten Canadian Elections & Unity

Global News warns AI deepfakes are blurring truth and fiction at increasing speed, with experts flagging Trump administration as a prime source targeting Canada amidst Alberta separatist talk. Officials express deep concern over disinformation, while researchers demand urgent training and skepticism as U.S. fakes flood X, Facebook, and TikTok.

Key Takeaways:

  • 📰 U.S. Disinfo Surge: Trump admin shares obvious deepfakes like AI Pope images, agents outputs on ditching truth.

  • 🗳️ Election Peril: Canada braces for AI interference rivaling Russia and China, with MPs briefed but leadership absent.

  • 📱 Platform Woes: The U.S.-owned social medias denies the statements, with 83% of fakes spread by Canadian people.

  • ⚠️ Unity Risk: Alberta independence hype +"51st state" memes ignites fear of state-backed false info potential worldwar.​

Rapid Fire Resources

AI Voice Input

Voice AI interaction platform.

Enterprise Decision AI

Enterprise AI analytics platform.

Personal Email AI

Personal AI email assistant.

No-Dode AI Agents.

AI agent building platform.

📊 Take This Edition’s Poll:

How was today's edition?

Rate this newsletter.

Login or Subscribe to participate in polls.

Why It Matters

These changes are needing strategic decisions. Wield Firefly for assets, benchmark agents rigorously, and guarding against the deepfake mess in social media platforms.

Operate a single high-impact workflow, tracking real-time upgrades, and enforcing human oversight. Only use what proves to be credible and reliable.

Until next time,

P.S. Interested in reaching our audience? You can sponsor our newsletter here.

How was today's edition?

Rate this newsletter.

Login or Subscribe to participate in polls.