This website uses cookies

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

Hey {{first_name|Conscious Church Fam}} 👋

Another big one. OpenAI just rebuilt its entire desktop app around getting real work done, Anthropic found a part of Claude's brain that nobody put there, and Meta dropped a free image model straight into your group chats.

Here's what we're covering:

  • 💼 OpenAI sends ChatGPT to work

  • 🧠 The workspace nobody built into Claude

  • 🎨 Meta's image AI lands in WhatsApp

Let's dive in 👇

✍️ Josh's Musings

First, the football. Is it coming home?…

For everyone who emailed after last week's England v Mexico match: yes, it was a quiet breakfast in our house. Maha was gutted in the moment, though she has since decided she is happy for England, and was really into the game on Saturday night. The World Cup rolls on, the heat rolls on, and so do we.

Second, something that genuinely surprised me this week with AI.

Maha has been planning next year's home education for the past few weeks. We love the approach we use, but it is very much curate-it-yourself, and most of the off-the-shelf versions are heavily Americanised, which is odd for something originally built on the work of an English Christian woman over a century ago.

So one evening I fed Claude Fable (the NEW LATEST BEST MODEL yada yada - it is genuinley brilliant though) the FREE open-source materials we already use, asked it to restructure them into a proper week's plan, parent guides and some specific added elements with an English flavour (eg History, Geography etc), and went to bed/fell asleep midway through prompting and it somehow got through to it.

By breakfast Maha was reading through printed PDFs, complete with maps and graphics, and the only fault in the whole pack was one minor design flaw.

Facts right, structure right, ready to use.

Now to clarify, the content was there so it wasn’t inventing or fabricating. It was just putting it an easy to use package. We will still buy a curriculum this year, but we will supplement it with custom units built exactly this way.

This is where I feel that for people and even Church’s AI is most helpful if you are using it.

Adding that visual, helpful layer on top of the human made, human done, content/data etc.

Tools that utilise the tools of today without stripping out the heart and rendering it soulless.

What about you? What has AI quietly done in your life/work recently that surprised you?

Stay Conscious, Stay Wild.
Josh

Image: OpenAI | The Conscious Church

OpenAI has launched GPT-5.6 and rebuilt its desktop app around it. Codex, its coding agent, is now merged into the main ChatGPT app, and a new agent called ChatGPT Work takes on bigger tasks for ordinary users. It is OpenAI's answer to Claude Cowork, and it turns ChatGPT from a chatbot into something closer to a colleague.

The Details:

  • GPT-5.6 comes in three tiers: Sol (flagship), Terra (everyday) and Luna (fast and cheap), with Sol priced at $5/$30 per million tokens, the same as GPT-5.5

  • The Codex desktop app is merging into a revamped ChatGPT app with a built-in browser and computer control, so coding, chat and agent work all live in one place

  • ChatGPT Work runs larger multi-step tasks on web, desktop and mobile, with adjustable effort levels depending on the job

  • Sol sits slightly below Anthropic's Fable on intelligence benchmarks but leads on agentic coding, and it even post-trained its smaller sibling Luna autonomously

Conscious Take:

Sol does not beat Fable on the benchmarks, but it is cheaper and it lives in the app you already have open. Most people will not choose the smartest model, they will choose the nearest one. That is the entire strategy.

Image: Anthropic | The Conscious Church

Anthropic researchers have found an internal structure inside Claude they are calling "J-space", a kind of mental notepad the model uses while it thinks. Here is the strange part: nobody built it. It emerged on its own during training, and it looks remarkably like the "global workspace" theory neuroscientists use to describe human working memory.

The Details:

  • J-space holds the concepts Claude is actively juggling mid-thought, separate from the visible chain-of-thought text it writes out

  • When researchers directly edited it (swapping the concept "spider" for "ant"), Claude's answers changed to match

  • Deleting it left basic abilities intact but severely damaged multi-step reasoning

  • Anthropic is explicit that this is not evidence of consciousness, but the parallel with human working memory is what makes the finding remarkable

Conscious Take:

Nobody designed this structure, it grew. We are now discovering things inside these models rather than engineering them, and whatever you make of the consciousness question, that alone should keep us humble about what we are building.

Image: Meta | The Conscious Church

Meta has released Muse Image, a new image model from its Superintelligence Labs, and it went straight to second place on the global leaderboards. The bigger story is where it lives: rolling out free inside Meta AI, Instagram and WhatsApp, with Facebook and Messenger next. Billions of people are about to have serious image generation sitting in their pocket.

The Details:

  • Muse Image ranks #2 on Arena's text-to-image and editing leaderboards, behind only OpenAI's GPT Image 2

  • It can search the web, use tools and edit its own output as it works

  • Free across Meta AI, Instagram and WhatsApp, with Facebook, Messenger and Meta's ad tools to follow

  • A companion model, Muse Video, was previewed at #3 on the video leaderboard

Conscious Take:

Free, genuinely good image generation inside WhatsApp means that you don’t even need to leave the app to generate AI Slop ‘sigh’. Memeability rating 10/10.. everything else… I’ll hold fire

📬 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.

Keep Reading