
Hey {{first_name|Conscious Church Fam}} 👋
Grok Bot can now sit on hold and haggle over your bills for you, American voters have had enough of data centres, and ChatGPT has learned to make GIFs. Not a bad week if you hate phone calls.
Here's what we're covering:
Grok Bot starts making your phone calls
Voters say no to data centres
ChatGPT's images start to move
Let's dive in 👇
✍️ Josh's Musings
I saw this tweet this week and was intrigued.
An AI robot that stabs fish in the brain. It's from a Californian startup called Shinkei, with Peter Thiel's Founders Fund among the backers.
The machine is the size of a fridge and sits on the deck of a fishing boat. A fish goes in, the AI works out the species and where the brain is, spikes it, cuts the gills and drops it into ice. It takes about seven seconds. I do wonder what Peter would have made of it. The fisherman, not Thiel.
It's based on ikejime, an old Japanese technique. Normally a lot of fish just suffocate on deck, which is grim and apparently ruins the meat, because they get stressed. Ikejime is much quicker, but it takes a skilled person, so it's mostly stayed rare and expensive outside Japan. Shinkei reckons its fish can last up to three times longer, though that's their claim. As one person on X put it, "Wow fish are really lucky that suffocating to death makes them taste worse."
I like it because it's so specific. It takes one skilled job that only a few people could do and makes it available on far more boats. Someone else on X said we've got too many people building agents and not enough using software to improve fish. Fair.
It’s fun seeing AI improve industries, not just the creative sector that I have my head stuck in each week.
Stay Wild, Stay Conscious.
Josh
Elon Musk's SpaceXSI has given Grok Bot, its AI agent, the ability to make phone calls. You tell it who to ring and what you want, and it sits on hold, talks to whoever picks up and reports back. Early users are already getting it to haggle down their bills.
The Details:
It's for paying users and only works with US numbers for now.
One early tester let it spend about 25 minutes on the phone to Comcast. It turned down three upsells and got the bill from $84 to $45 a month, on a faster plan.
You follow a live transcript while it talks. If the rep asks something it doesn't know, you get a message and type the answer for it to pass on.
The bot has to say it's a virtual assistant, and the terms reportedly make you the "caller of record", so you're the one responsible for the call.
Conscious Take:
Sitting on hold is exactly the sort of job I'd happily hand over. Especially Wifi/TV upgrades!! But it does feel a bit cheeky especially if they other end of the line is a human operator.
Data centres turned into one of the big issues at the US midterms. Towns and counties across the country got to vote directly on whether to limit them, and the answer was pretty much yes everywhere. Democrats also won the House, promising tougher oversight of the tech companies.
The Details:
Limits on data centres have won 44 of the 45 local ballot measures with complete counts, with two still being counted. All of it is unofficial until certified.
In the typical measure, the anti-data centre side took 77% of the vote.
In NBC's exit poll, 56% said AI and data centres were a crucial or major factor in how they voted.
Goldman Sachs reckons the states likely to tighten the rules hold about 45% of planned US capacity, though it expects any pauses to be short-lived.
Conscious Take:
Most of the AI story so far has been decided by a handful of companies and governments. This is the first time normal people have had a direct vote on any of it, and they said no by a lot. OpenAI's reported plan is to start calling them "AI compute campuses", which I can't see fixing it.
OpenAI has released Images 2.6, its new image model, and the fun bit is GIFs. Type an idea, or give it a photo or a rough sketch, and it'll turn it into a looping animation.
The Details:
It's rolling out globally across all ChatGPT and Codex subscription tiers, on desktop, mobile and web.
Images generate 50% faster than before. Flare, the lighter version for developers, can manage one in about six seconds on low quality.
Characters, products and brand details should stay more consistent as you make edits.
OpenAI says people now make 4 billion images a week with its tools.
Conscious Take:
I can see this being really handy for churches, a little looping graphic for a kids' talk or a sermon series without needing a favour from the one person who knows After Effects. It isn't the big model upgrade people are waiting for. I reckon more people will actually use it though.
📬 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
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.



