Eight years later

Hello again my friend,

When I started writing Liight Work, I covered the future of work. Careers, content, AI tools, and their impact on industries / the people in them (customers, management, or CEOs). The question was how all of it would change your job or your business.

But the stories are getting stranger, the outcomes are often absurd, and (in many cases), we’ve reached that future.

I remember sitting to write about this cool new tool called ChatGPT when it first released. When Midjourney made images out of nothing (and they were great). When AI learned how to code little apps.

Now, all of this is flipping the world on its head. The technology is so powerful the government is getting involved in recalling it (temporarily). The same companies building AI are funding the programs for displaced workers. Brain chips are restoring people’s ability to speak after it was lost, and dog collars can translate barks into human speech.

That is an insane paragraph.

But it is real.

And that's the point this week. The last two years moved us 8 years into the future, and most people are still reacting to each story like it's a novelty instead of realizing the timeline changed.

Let's lock in.

The biggest story this week might be the least flashy one.

OpenAI staggered the release of GPT-5.6 after a request from the U.S. government. The model went first to a small group of trusted partners, with government involvement in the rollout. Yesterday, the U.S. lifted the export controls it had placed on Anthropic's Fable and Mythos models over concerns about cyber misuse.

But why? Why would the US government do this?

Because the new models can produce code that exploits security vulnerabilities. Amazon's researchers demonstrated it, and that report drove the recall. That’s who reported the issue, and the big headline driving the rollback. But think about what it implies. What could it build for people? Could it make weapons? It’s kinda scary.

A recall might sound like boring policy, but it’s actually a massive shift.

AI is being treated less like a normal app and more like infrastructure. Not exactly weapons, not exactly cybersecurity, not exactly telecom, but somewhere in that zone where governments start asking who gets access, what the risks are, and whether the most powerful version should be available to everyone at the same time. Dictatorships are often criticized for taking away the people’s access to internet. What happens when governments decide to take away people’s extra intelligence (but keep it all for themselves?).

There are two major takeaways for the rest of us.

What do I mean by open source?

Claude and ChatGPT are owned by private companies. They house them and lease them to us. Open source models are the ones we can also house ourselves (they typically can’t be taken away at that point, at least not as easily).

Open source may be the future, but I’ve tried Fable 5, and it’s amazing. Excited to try GPT-5.6.

This is like when tobacco companies sponsored cancer research.

Axios reported that Anthropic and OpenAI are joining a $500 million effort called Raise Us, led by former Commerce Secretary Gina Raimondo and former Indiana Governor Eric Holcomb, to help workers and states prepare for an AI economy. The group hopes to raise $1 billion in total, with anchor partners including the OpenAI Foundation, Anthropic, Microsoft, Amazon, and Bank of America.

This is one of those stories where the existence of the thing is more interesting than the press release.

The companies building the technology that could disrupt huge parts of the labor market are now helping fund the transition plan for the people who may be disrupted by it.

That doesn't mean everyone loses their job tomorrow, these things move unevenly. Tech feels it first, then large companies, then small businesses, then everyone else once the tools get cheaper and easier to use.

But you do not create a $500 million workforce transition effort unless serious people believe transition is actually coming.

That's the part I keep sitting with. Working in tech makes me feel like I’m a few moves away from getting slapped with something. Maybe not job loss, but something, and it makes me uneasy.

A year ago, the career conversation was about how to use AI to get ahead. Learn the tools. Get faster. Automate the boring parts. Become the person at work who knows how to use the new thing. Awesome. That advice still matters. But the conversation is getting bigger. It's no longer just "how do I use AI to be more productive?" It's "what happens to the labor market when millions of people use AI to be more productive at the same time?"

Those are very different questions. If you’re not at least trying to use AI, please do.

Business Insider reported that Paradromics, a Neuralink competitor, implanted its Connexus brain chip into a human patient for the first time. The patient is a Michigan woman with a motor neuron disease that affects her speech. The goal is to record the brain signals associated with speech and eventually translate them into text or a synthesized voice.

I know. Brain chip. Human patient. Restore speech.

But here's why I take it seriously: the first use case is not some fantasy about uploading your brain or becoming superhuman. It's much more human than that. Someone can't speak properly, and the technology is trying to help her communicate again.

That's usually how strange technology becomes real. Not through the most extreme version, but through the most obvious human need. If you can help someone speak, move, hear, see, or regain some piece of independence, the conversation changes. It stops being weird and starts being useful. And once something becomes useful, patients, doctors, regulators, and investors all start pulling it forward.

Brain-computer interfaces are still early. They're risky. They're medical devices, not consumer gadgets. But they're no longer just something people argue about on the internet because Elon Musk tweeted about Neuralink.

And then, because the future wants to give us some sliver of hope, we have this.

A company called Traini is selling what it calls a real-time human-to-dog conversational collar. The claim is that the collar can translate human speech into dog sounds through an app, and it doubles as a smart collar with GPS and health monitoring. The accuracy claims deserve plenty of skepticism.

Do I think we're about to have deep philosophical conversations with dogs?

No. I’m a dog guy, I love dogs, but not yet. Pay attention anyway, because the category itself is the story.

We already track steps, sleep, glucose, heart rate, spending, workouts, and screen time for ourselves. Of course that logic is going to move to pets. People treat pets like family and already spend absurd amounts of money on them. If a collar can tell you your dog is stressed, sick, lost, or acting differently than usual, people will buy it.

The talking dog is the headline. The real business is monitoring, interpretation, and attachment. Even if all it can accurately tell you is 'happy,' 'sad,' or 'scared'" that’s still incredible and I’d buy it.

When I started this newsletter, the future mostly felt like a work story. Better tools. Faster content. New career paths. New ways for one person to do more.

Now it's showing up in government policy, labor programs, hospitals, and so much more. If we can get this far in a few short years, what do the next ten look like?

I can’t say for sure. I know some of this will remain overhyped, some will fail, and regulation is here to stay. Governments will continue stepping in front of models, and entire nations will be more aggressive in their strategies.

The future never arrives clean.

You know how there’s always a big moment before everything changes? I feel like we’re at the part before that right now. I think people under-rate the impact AI will have on humanity (yes, even after everything you’re seeing, people don’t give it enough credit).

And so yes, I think we’ll see more change, faster than ever.

Thanks, as always, for reading.

Darwin