This website uses cookies

Read our Privacy policy and Terms of use for more information.

Hey {{first_name|Conscious Church Fam}} 👋

Morning! This week's AI news went from wonderfully weird to genuinely important, sometimes within the same story.

Here's what we're covering:

  • Claude designing proteins better than professional drug discovery teams

  • A doctor solving a 22-year maths problem with ChatGPT

  • Hollywood's first real truce with an AI video company

  • Why I've just hard-launched Voco (and it's not just for translation)

Let's dive in 👇

✍️ Josh's Musings

A guy called Jonathan Wilke vibe-coded a website this week called outbid.lol. Three hours, start to finish. It's a leaderboard where you pay to be top (and promote your website). No algorithm, no vote, no claim the best product wins. Just whoever pays most sits at the top.

$139,000 in 65 hours. Over a million visitors. He turned down a six figure offer to buy it off him.

It’s been all over my X feed the past few days. He launched it on a whim and it just blew up. Much like that million-dollar homepage site 20+ years ago.

Then I was chatting to my parents this week about their work. They are both excellent at what they do, hands-on, the actual doing part. What eats their week isn't the work itself. It's the admin after it. The notes, the write-ups, summarising what just happened so someone else can read it later.

Some of the tools they already use now have AI bolted on. So I went looking for something that wasn't just their phone propped on a desk. Something that could record, sync to the cloud, and summarise on its own.

Found a device called Pocket (not an ad or affiliate, but would happily test it haha). Screenless, just audio records, syncs when you're back near your phone, writes it up for you. Looks genuinely great, and you don’t need to keep remembering to open your phone and voice record.

Has anyone got one/tried it?

AI tools/products that just ‘work’ and stay out of the way and enable the bits that we as humans excel at are the way to go.

That's the whole reason I built Voco Translate, if I'm honest. Not a clever feature. Just getting the language barrier out of the way so the message actually lands.

I hard-launched it properly this week, after months of building and weeks of talking with Church’s. A few mentioned not having any international speakers, yet found the idea of those hard of hearing being able to follow along in English to be helpful, or to be able to add live captions to a video feed in the baby room.

No credit card, two minutes to set up, and if you're not sure, just hit reply and tell me about your church. → Voco.church (send it to your pastor, a friend, or a church you know).

Stay curious, stay conscious.

Josh

Image: Alextownsend.net | The Conscious Church

A Beijing neurosurgery resident just cracked Crouzeix's Conjecture, a matrix maths problem that had stumped mathematicians since 2004, using nothing but a self-taught grasp of advanced maths and a 16-hour ChatGPT session.

The Details:

  • Shanmu Jin found the problem through his own brain-ultrasound research, having taught himself advanced maths after training in geology and medicine

  • He ran GPT-5.6 Sol with no internet access, setting swarms of AI subagents against each other to stress-test the proof

  • The proof has already been checked by Cornell mathematician Alex Townsend, Anne Greenbaum, and the conjecture's original author, though formal peer review is still pending

  • It's one of a run of AI-assisted maths breakthroughs this year, alongside solved Erdos problems and progress on the Riemann hypothesis

Conscious Take:

What gets me is that Jin isn't even a mathematician. That being said, I’m sure he’s better than most. But it’s crazy now that the real skill in the imminent future will be those who can verify something is real/true/correct.

Image: Anthropic | The Conscious Church

Anthropic let Claude loose on a proper drug discovery task this week: designing protein binders against 15 different targets with almost no human input. It succeeded on 14 of them, at hit rates well above what human-led teams typically manage.

The Details:

  • Claude (Mythos Preview and Opus 4.8) ran the whole binder design process itself, choosing targets, generating candidates, and screening results, work that normally takes a specialist weeks per target

  • Hit rates landed between 22% and 35%, against a typical 10-15% in the field, with some designs binding more tightly than the best published results

  • Separately, Opus 5 read a lab's raw chemistry instrument files with no vendor software, matching the lab's own purity readings in under 20 minutes

  • The catch: this stays behind a research access programme for now, not something general-access Claude can do today

Conscious Take:

You know I love these AI stories as I’m sure many of you do. It’s exciting to see solutions that were possible, albeit unknown to us, now becoming real and tested in a short space of time.

Image: Ruairi Robinson (@RuairiRobinson on X) | The Conscious Church

Six months after the Motion Picture Association sent its first ever cease and desist to an AI company, ByteDance has agreed to build proper copyright guardrails into its video tools. It's the first formal truce between a major studio group and an AI video maker.

The Details:

  • The MPA (Disney, Netflix, Sony and co) accused ByteDance's Seedance tool of generating unauthorised clips of characters like SpongeBob, and copying scenes straight from Stranger Things

  • The complaint followed a viral deepfake clip that helped kick off the wider Hollywood versus AI video fight

  • ByteDance says the fix is baked into newer releases, Seedance 2.5 and Seedream 5.0 Pro, though neither side will say exactly what the guardrails are

  • Indie filmmakers already lean on Seedance because it undercuts rivals on price, so this deal affects a lot more than just the big studios

Conscious Take:

Studios spent years threatening lawsuits while AI video got good enough to actually worry about. This is probably a preview of how the rest of Hollywood handles this too. Especially now that the tools are at a level more than good enough to be a part of the creative arsenal.

Image: Voco | The Conscious Church

I properly hard-launched Voco this week: live sermon translation for churches in 200+ languages, no app and no hardware. Sign up takes about two minutes and there's no credit card involved.

The Details:

  • Live translation and captions in 200+ languages, delivered straight to a phone browser, nothing to download

  • It's not only for language translation. Churches can use it purely for live captions, for deaf and hard-of-hearing members or anyone who just finds it easier to read along

  • Free 7-day trial, no card required, and paid plans start from £6/week with no contract

  • Works with the setup you already have (a phone or laptop mic), plus OBS and ProPresenter support if you want it up on screen

Conscious Take:

This has been months of building, and I'm genuinely proud of how simple it is to try. If your church has anyone who doesn't speak English first, or anyone hard of hearing, I'd love for you to give it a go.

📬 One quick ask...

If this email has been helpful, would you forward it to one person this week who might be interested? Thanks :)

"Whatever you do, work at it with all your heart, as working for the Lord, not for human masters." : Colossians 3:23

Build with the tools. But build for the right reasons.

That's all for now

To help us make this an even better experience for you, we'd love to know your feedback from the email today.

Login or Subscribe to participate

Stay conscious,

Josh

P.S. If you liked this then please forward it on to someone you think would enjoy it. And if someone forwarded you this and you liked it, you can sign up here.