- AI For Humans - The Newsletter
- Posts
- đź”® Our AI Predictions for 2025 đź”®
đź”® Our AI Predictions for 2025 đź”®
Buckle up, because things are about to get weird
Today on AI For Humans The Newsletter!
We get the crystal balls out for 2025
Google just made podcasts interactive
And robotics developments continue to give us nightmares
Plus, our can’t-miss AI feature of the week!
Welcome back to the AI For Humans Newsletter!
We can all agree 2024 exceeded expectations, but 2025? We think it could be even more juicy 🤌
Freshly squeezed from Friday’s prediction video, here are 7 headlines we may see in 2025:
AI + Crypto Synergy:
Expect the crypto world to latch onto AI. We’ll see more “AI agents” or “AI personalities” tied to tokens and on-chain transactions—especially in the first half of the year.Code Generation Gets Fully “Cracked”:
AI coding assistants will improve so much that “one-shotting” entire websites or apps becomes possible. People with zero coding background will spin up surprisingly robust software.2025 = The Year of the Agent:
Autonomous AI systems that break big tasks into smaller ones (and complete them with minimal human supervision) will be everywhere—personal productivity, enterprise workflows, and beyond.
“Prompt to Hollywood” Approaches:
AI video tools (like Genesis) will merge 3D engines (Unity, Unreal) with realistic generative content. Studios and VFX houses might shrink their teams from hundreds of artists to just a handful.
New Frontier Models by Mid-Year:
With OpenAI’s O3 teased at the end of 2024, expect a flood of next-gen models from OpenAI, Meta, xAI, Anthropic, and others. Each might be only “25% better” than the last, but those steady improvements push us closer to AGI.
AI’s Role in Warfare:
Military tech is advancing quickly, and new AI capabilities will almost certainly spill into defense. Fully or semi-autonomous systems could see major breakthroughs.A Major “Mass Hallucination” Moment:
Hyper-realistic AI-generated videos could spark a large-scale misinformation event that fools the media and the public, at least briefly.
The one thing we’re 100% sure of is 2025 will bring more surprises than ever, and we can’t wait to share the journey with you.
See you next week as we return to our regularly scheduled programming!
-Kevin & Gavin
3 Things To Know
NotebookLM’s Podcast Feature Gets A BIG Upgrade
This one went under the radar during OpenAI’s shipping-spree: Remember the feature that generates a podcast about any subject? Now you can interrupt the hosts to join the conversation. In this example the author of a book joins a discussion about the book. Making static media an interactive experience? Sign us up.
A Straight Shot to Artificial SUPER Intelligence?
Google AI product lead Logan Kilpatrick joined the chorus of a growing number who believe scaling inference will get us to artificial super intelligence. ASI is the threshold at which AI exceeds human capabilities across all domains, and inference is compute juice spent letting a model reason through a problem. In the past it seemed a number of breakthroughs would be required to reach ASI, but it appears we could get there by simply throwing more compute at today’s approach.
This Robot đź‘€
We’re not sure what’s happening, but, [insert thought leadership here].
We đź’› This - Sora Loop
OpenAI’s Sora may not be the more realistic video model, but it comes packed with some incredibly powerful features. Among them, the ability to create a perfect loop. So we took this little video created with Kling, and used Sora to loop it.
Simply trim your video down, indicate how much time should be added to complete the loop, 2, 4 or 6 seconds, and hit generate.
Here’s the output. Sora nailed it on the first attempt, and it really is seamless!
Are you a creative or brand looking to go deeper with AI?
Join our community of collaborative creators on the AI4H Discord
Get exclusive access to all things AI4H on our Patreon
If you’re an org, consider booking Kevin & Gavin for your next event!