In partnership with
Today on AI For Humans:
Preparing For More AI Advancements
Is This Finally DeepSeek V4 Week?
Plus, Baseball Players Are Vibecoding
Welcome to the AI For Humans newsletter!
How are you doing? Feeling ok?
It’s worth a basic human check-in after this truly unprecedented and crazy week.
We’ve finally come to the time we’ve discussed on the podcast for years, where the issues surrounding AI and the cutting edge models are directly intertwining with the worlds of politics and business.
Now that is a man who looks like he’s been through a lot this week.
ICYMI, there was a massive dustup between Anthropic, the company behind the Claude chatbot, and the United States Department of War.
This has, for now, led to Anthropic being declared a ‘supply chain risk’ by the US government and, somehow, to OpenAI’s Sam Altman swooping in to do the deal Anthropic would not.
Oh, and also Claude is now #1 in the app store and has Katy Perry’s support.
We’re not going to re-hash this story (many other media outlets have taken it on) but will try to discern a few takeaways from the barrage of news this week and leave you a bit more prepared for the rest of the year.
First and foremost, as much as we might want it to, none of this is slowing down…
Please support AI For Humans by learning about our sponsors below:
Want to get the most out of ChatGPT?
ChatGPT is a superpower if you know how to use it correctly.
Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.
Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.
The Speed of Change Is Getting Ever Faster…
If the week we just went through feels like a lot, well, that’s because it was.
Here’s just a sample of the AI-related headlines we saw:
Viral Doomsday Report Lays Bare Wall Street’s Deep Anxiety About AI Future (Citrini Substack Fallout)
Ok, that last one is a little plug.
But it ties into this idea that things are changing faster than we’ve ever seen before.
I made that website. Me. From start to finish, I conceived of it, designed it, ‘coded’ it, handled a semi-complicated back-end database, debugged it and deployed it to the web.
Ofc, I did all that with the help of Claude Code but even a few months ago I couldn’t have conceived it would be possible.
My point here is yes, last week felt crazy but it’s not unusual anymore.
A few weeks ago here, I explained AI takeoff and I urge you to read that if you haven’t.
Change in this space is coming faster than ever. And it’s gonna keep going.
This feeling you’re having… it’s likely not going to end.
But With Great Change… Comes Great Opportunity
There’s a version of myself where I feel all this happening and I just want to turtle inside.
Just let the change kind of wash over me and, like a good friend of mine has planned, go find a dive bar to run. Kind of a nice life.
HOWEVER, there’s another part of me that says ‘Huh, maybe I can remake the AI For Humans website, make it WAY better and save myself $20 bucks a month’.
That part of me won this week.
It’s overwhelming the amount of new tools being shipped or new AI models coming out or end-of-the-world-scenarios being thrown around but… you can find your own personal signal in the noise.
The thing I’m thinking a lot about these days is how AI agents are likely to reshape the entire way we find and distribute content to humans.
I’ve thought a lot about that particular issue over my career and I’m letting that problem noodle in my brain.
It’s an area I have specific domain expertise in and, as long as I trust my instincts, something will probably come out of it.
A few scrolls down in this newsletter, you’ll see a lawyer who’s doing that for his business with Claude. Or a baseball player using AI coding tools to map pitch data in a way that’s truly useful to him and others.
Or you can, like popular AI dev & YouTuber Theo, decide to take on Adobe’s Frame.io and create an entirely new and open-sourced version just because you know what a terrible and pain-in-the-ass product it is.
Theo saw a gap and just... built the thing.
The point here is…
When things change fast, you can too. You just need to be ready for it.
Love this newsletter? Forward it to one curious friend. They can join in one click.
But… It’s All a Bit Much Isn’t It?
Yes. Yes, it is a bit much.
And, to be completely and totally honest, it’s such a bit much that in five years everything that we’re learning right now might be obsolete.
But part of living in these weird times is to get yourself used to this pace of change.
And, more importantly, to not be surprised by it.
If you can get to that place, where you’re not only expecting the change but starting to anticipate it and even get excited for it, you’ll be fine.
See y’all next Friday for a new episode!
-Gavin
This week’s AI For Humans: New Nano Banana 2 Is Not Bad But Seedance 2.0 is Great👇
3 Things To Know About AI Today
Speaking of Big Change: DeepSeek V4 Finally This Week?
So… I’ve been guilty here of buying into a bunch of rumors around the newest AI model from Chinese start-up DeepSeek.
I’m sorry… but here we are again.
This time, we’ve finally gotten real info from a credible source and it looks like (according to the Financial Times) that we’ll see V4 this week.
The FT article is behind a paywall… but there are ways around it.
Some pretty compelling factoids from this:
V4 is expected to be released this week ahead of China’s ‘Two Sessions’ meeting where the country communicates macro-economic goals.
It is a multi-model model and (supposedly) will produce text, image & video. Word is that it is very good.
While the FT article discusses the fact that they worked exclusively with Chinese chip makers on the new model, Reuters reported this week that it was illegally trained on NVIDIA’s Blackwell chip.
If (and this is a big if still) Deepseek v4 does appear this week and takes over Opus 4.6 as the current SOTA AI model, get ready for even more market chaos.
Step-By-Step On How To Copyright AI Video Work
It's an ongoing myth that work containing output of an AI video or text generator cannot be copyrighted.
Yes, there was a ruling that a purely AI creation cannot receive a copyright designation but literally any human input on the output after the fact (or on the designs going into the machine in the first place) can show human creation.
Wikipedia has a full page dedicated to the subject which is worth reading. Don’t Google it because the SEO demons have come out to play there.
That doesn’t mean it’s as simple and direct as all that.
Friend of the show/newsletter PJ Accetturo walked through a recent campaign he ran for his AI agency and how he made sure that human beings were used at nearly every step along the way.
The Claude Native Law Firm
I come from a family of lawyers (father, grandfather, uncle, two cousins, etc) but I personally never loved the idea of doing that work.
The idea of standing up in front of a jury and arguing my case was compelling but I saw first hand how much of law is thumbing through cases and dealing with endless legal forms.
Not my cup of tea.
Legal AI start-ups like Harvey have revolutionized a ton of this brain-numbing work and have gotten huge valuations in the process.
But what if… all you needed was Claude?
This is yet another example of how tenuous the AI start-up world is.
When the state-of-the-art is progressing so fast, the large models level up and are suddenly able to do things that only specialized models could do before.
But, again, there is opportunity in one person figuring out how to do this themselves.
Cursor is another specialized start-up (coding) that is in the middle of an identity crisis. After all, why go to Cursor when you can just do it all inside Claude Code?
This was an interesting post from founder Michael Truell about how they’re approaching the next generation of their company.
We 💛 This: Claude Coded Pitch Analysis From A Non-Coding Baseball Player
As we mentioned up top, all humans have some sort of domain expertise.
You have something in your life that you know better than most people and that could be your secret sauce when it comes to making something incredibly valuable.
For instance, maybe you’re a baseball pitcher and understand that while there’s a lot of data in baseball, there’s one specific aspect of it that’s missing.
Robert isn’t uniquely talented, he’s just early.
He took his domain expertise in pitching and spotted a specific gap in the data the industry needed.
What is your specific domain expertise? How can you vibecode your way to a product?
Now is the time, y’all. Start mining the sorts of things that only you know.
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!