Kia ora {{ First name | friend }},
We’re exploring the light and shadow of technology, with a focus on efficient AI workflows to free up your time.
This week’s edition features:
Blotato social media scheduler
OpenAI GPT-5.1 and their court case
Anthropic’s new announcement
Releases from Canva and ElevenLabs
Gmail watch-out
Reflections on AI competency gap
Thanks for being here.
Thanks to my friend Claire for this one.
This tool looks like a master stroke for social content creators. Ideate, generate and distribute all in one place. I haven’t given it a go yet, but at face-value, it looks next level. One of the things that is most attractive to me is the ability to plug it in to other existing AI workflows, from Airtable, Zapier, N8n, Make etc.
I’ve started working on a workflow such as the below
Input post idea by messaging a Telegram bot, or forwarding myself an email
(Post ideas on Blotato can also come from podcast files, PDF’s, text files, or anything else really)
For certain contexts, such as news, I may also look to set up an n8n web scraper to get the top daily news on a subject, and store those article summaries in Airtable
Post idea originally stored in Airtable
Airtable automation triggers Blotato post idea
Blotato creates content
Perhaps Canva required too
Perhaps trigger an external content automation such as HeyGen for video or Nano Banana for imagery
Blotato reformat original content into a series of posts for IG, X and Tiktok
Schedule posts through Blotato
Iterate and repeat
There are definitely still some human-in-the-loop gaps in the above flow and will take a while to get it right.
Yes, this sort of workflow is bound to add to the mountain of AI slop on the internet. For sure. It’s not necessarily good for very community-focused brands/projects, or ones that value tailored, high-quality media for their audience. If you’re trying to scale a project fast, efficiently, and don’t care so much about brand rep, this is more suitable for you.
PS: Don’t drink the Kool-Aid
It’s worth noting this wider point here - which is around the AI sensationalism around “Automate your whole business” - every joker on IG is selling the PDF/course/workflow with a photo of the beachside laptop workstation. It’s possible, yes, and a lot of it is hype. Here’s the shadow I perceive
The offer/tool is usually technical AF and takes ages to build, definitely not beachside chill vibes, at least for the first while
I’ve noticed that highly-automated online offers often lack the human touch and substance.
We’ll continue to smell the BS and there’ll be a growing list of people who don’t wanna drink the AI Kool-Aid and actually crave real human content
The hype and mental indulgence of the whole AI thing can be super overstimulating for the nervous system. If you haven’t got an equal balance of embodiment tools, nature time, real human connection, it’s a dangerous path to walk.
That said, the cat is out of the bag, and we have a chance now to give it a shot. AI gonna continue doing weird stuff. Let’s just have a splash around and see what happens.
Watch Out: Gemini AI can now read all of your emails
On October 10, 2025, Google quietly switched on new Gemini-powered “smart features” for Gmail, Chat, and Meet for many users by default. These features read and analyze the content of your emails and attachments to drive things like summaries, suggested replies, and a more detailed advertising profile based on your private messages and documents.
Google frames this as personalization across Gmail and Google Workspace, linking it to conveniences such as Smart Compose, automatic inbox categories, trip cards, and Gemini drafting features.
In practice, it means your inbox now feeds two systems at once, one that tries to save you time and another that learns your habits, purchases, travel, and relationships to sharpen targeting and improve AI models.
This seems like efficiency on the surface (it is), however the rabbit hole goes deeper, where now Gemini is being trained on your data, and no doubt Google will leverage that data for even more targeted ads and ‘personalisation’.
How to change it
On desktop, open Gmail, click the gear in the top right, then “See all settings.” In the General tab, scroll to “Smart features and personalization” and turn that off. Then find the Google Workspace smart features section, click “Manage Workspace,” and switch off both “Smart features in Google Workspace” and “Smart features in other Google products.” Repeat a quick check in the Gmail mobile app settings for the same smart feature options. After that, Gemini has far less access to use your inbox for training or ads.
What else to look out for in future of AI
Expect more products to turn on AI helpers by default and hide the real tradeoffs behind friendly language like “personalization” or “smart suggestions.” Watch for any setting that says your data is used to “improve services” or “help train models,” especially in email, docs, calendars, calls, and home devices. Build a simple routine once a quarter where you open your main tools, scan privacy and AI sections, set everything to the most private option first, then only switch features back on when they save clear time or stress.
PS, you might like to shift to Proton Mail to avoid more infiltration in your emails :)
Here’s a few interesting technology updates:
OpenAI released GPT-5.1, a major update designed to improve reasoning, conversation stability and overall “human warmth” compared to GPT-5. Read more.
Canva dropped their “biggest produce release yet”, allowing on-brand image variants from language prompts, deeper edits, and much more accuracy. I’m curious to have a play with this over the next week, because on-brand imagery is still one of the big gaps in AI marketing. Watch the release video.
Anthropic announced a massive $50B investment in U.S. AI infrastructure, building out the company’s first custom data centers in Texas and New York starting in 2026.
OpenAI and New York Times enter the battle arena, where NYT is asking for 20 million anonymous user chats as part of an ongoing copyright lawsuit. OpenAI is calling the request an “invasion of user privacy'“. If OpenAI is actually respecting user privacy in other areas of their business remains to be seen… Read more.
OpenAI released their view on AI progress, as we edge closer to “superintelligence”, as it gets smarter, more efficiency, and cheaper to run. Very interesting article. Here’s a key quote from the article “The gap between how most people are using AI and what AI is presently capable of is immense.” (More on this topic below). Have a read.
ElevenLabs released “real-time” transcription with Scribe V2, supposedly the most accurate speech-to-text out there. The benefit in this is their API flexibility which is allowing people to build speech-to-text into multi-step workflows (e.g. upload a YouTube video, have it immediately transcribed, chapterised, short transcripts and social captions). Learn more.
Vibe Hacking has reached a new level, when Agentic AI hackers manipulated Claude. Allegedly a Chinese state-supported hacking group was able to automate 90% of hacking efforts with AI agents used with unprecedented skill. Anthropic apparently blocked the hackers and shut down the operation. You can read more here.
Exploration: “The gap between how most people are using AI and what AI is presently capable of is immense.”
This is a quote from the OpenAI article, above, where they shared their research on AI progress. Read the article here.
I feel a lot of resonance with this statement. Most people I speak to about AI are using it for daily tasks (how to do this, write this email, solve this challenge). Some are using it for deeper more intimate questions (therapist/coach, personal development etc), which I have said before and will say again, this is a very dangerous path in terms of your data sovereignty. I would advise stop engaging in intimate and personal conversations with ChatGPT (or any other LLM’s you’re not sure of their privacy policy) to keep your data safe.
Anyway, I digress, although this is important context.
The point is this: Intelligent AI is not coming “one day”, but today’s AI is already far ahead of how most humans direct it. Most of us general internet-punters are capable of much more than we think with AI, and I’m keen to unlock more potential by closing the gap.
Not “AI is coming one day”, but “today’s AI is already far ahead of how most humans direct it.”
So the question for you:
How do you move from casual use to serious leverage?
In my experience, it’s been through the following 3 key insights.
Build AI with AI
Treat AI as your thinking partner when you create workflows. Instead of staring at a blank spreadsheet or tool, you explain your goal, share data or screenshots, and ask AI to design the steps, formulas, and prompts. You lean on reasoning models to fill gaps in your knowledge, then you review and tweak the output. This closes the usage gap because you stop waiting until you are an expert, and you use AI itself to get you there faster. Sometimes it’ll hallucinate, but still faster than doing the manual work. Strongly recommend using GPT-5 Thinking or other reasoning models for this kind of work.
Expand, reflect and reiterate with multi step context building
A quick leverage hack is just to stop treating AI like a one question tool. You start with a rough idea or problem, ask it to ask you clarifying questions, then feed it examples, notes, or documents. You get a first draft, ask for critique, request a second and third version, and compare. This loop of expansion and reflection builds richer context over time, so the model mirrors your situation more closely and gives more useful answers. Caveat: be aware of the endless loop of iteration. At some point you’ve gotta take off the training wheels and hit the jumps yourself (i.e polish it and send it).
Focus on principles instead of tools
Stop chasing every new app and start asking better questions about your work. What decision do you need help with? What steps are you repeating every week? Where are you slow or inconsistent? When solving these, start small, easy, and effective. Just automate one small step, then the next, then the next. The gaps will reveal themselves later. Every launch of a new tool is distraction wrapped in nice packaging and an expensive monthly sub. Tools will always change but not all of them will get you results.
Leverage of this tech is more about principles, ideas and creativity, than it is about prompts, tools and models.
{{ First name | friend }}, thanks for reading.
If you found this useful, please reply and let me know what you enjoyed, and share this with a mate who you think could benefit from it. :)
Stay human,
Billy