MKingscott

A geeky software developer and #aiart enthusiast.

Buy my #ChatGPT+ #Midjourney guide here:

[MKingscott @M_Kingscott71]

#AIArtwork #AIart

I've written a 46-page guide on using #ChatGPT for #Midjourney titled "How To Programme ChatGPT For Amazing Midjourney Prompts".

It's now for sale on Gumroad for £2.99 here: https://t.co/E9ysL14s2n

Some images created via the process taught in the guide:

[MKingscott @M_Kingscott71]

Nice sunset (not AI)

[MKingscott @M_Kingscott71]

@O4tcake @greg16676935420 @elonmusk @grok Good one 👍

[MKingscott @M_Kingscott71]

@greg16676935420 @elonmusk @grok Eleventy trillion 👍

[MKingscott @M_Kingscott71]

@elonmusk Thats, like, a trillion dollars!

[MKingscott @M_Kingscott71]

@NEXUS_TO_NOVA Coolio!

[NΞXUS @NEXUS_TO_NOVA]

How to create martial arts videos with Seedance 2.0!

Prompt below👇

[MKingscott @M_Kingscott71]

@FynCas AI

[Zoomer @ZoomerHistorian]

We need to make the Nuremberg Trials look like a picnic. No one involved must escape the death penalty.

Heads must roll.

[MKingscott @M_Kingscott71]

@AverageJayX @LBC I doubt it's that detailed.

[MKingscott @M_Kingscott71]

@LBC Labour are censorious bastards, and should be treated as such. They are the modern-day stasi.

[Heather Cooper @HBCoop_]

Midjourney v8.1 --sref 7126546515

[Ruben Hassid @rubenhassid]

The slowest way to use Claude is to keep prompting it.

This (one-time) setup makes Claude work like you:

Cowork

☑ Chat forgets you. Cowork remembers:
Chat starts from zero every time.
Cowork reads your files before it answers.
↳ Download the Claude app, get Pro, open Cowork.

The Cowork folders

☑ A folder beats a clever prompt:
Make one folder called "Claude Cowork."
Add 3 subfolders: about-me, outputs, templates.
↳ Point Cowork to it. This is now Claude's brain.

The settings

☑ The defaults are why your answers feel generic:
Select Opus 4.8 for hard tasks.
Turn on Thinking & Effort = High.
↳ Block 20 minutes and set it up once.

The files

☑ Three sharp files beat 50 messy ones:
Open a fresh session. Ask Claude to interview you.
Build your about-me, anti-ai-writing-style, and my-company files.
Trim each one under 2,000 tokens. If a line doesn't change Claude, cut it.
↳ Don't write them from scratch. You can download mine at https://t.co/psB7XxAv8w.

The habits

☑ The setup is useless if you keep your old habits:
Talk, don't type. Install Wispr Flow (it's free).
End prompts with "ask me questions."
Call out the mistake. Don't accept it.
↳ Start a new session every 20 messages.

The system

☑ Most people quit before they build the system:
Save your best work as a template.
Reuse it instead of retyping it.
↳ One folder. 3 files. Claude works like you, forever.

Copy my folder & download my 3 personal .md files:

Step 1: Subscribe for free at https://t.co/psB7XxAv8w.
Step 2: You will have two choices: free or paid.
Step 3: Choose the free tier. Don't pay for anything.
Step 4: Open your welcoming email. Reply to it.
Step 5: Trace the Notion link. Open '.md files' folder.
Step 6: Access my entire folder + 3 files template.
Step 7: Send this image to your team's channel.
Step 8: Read 2x newsletter per week (for free).
Step 9: Become the "AI guy" at work, forever.

[MKingscott @M_Kingscott71]

@manairport I picked my son and his wife up from Terminal 2 on the 30th May, forgot to pay for parking in the barrier-less car park, and I still haven't received notice of a fine - what's going on? I was there for under an hour but I think you still pay for that?

[MKingscott @M_Kingscott71]

@Bogzabs96 REVIEWS

[Bogdan | Ad Creatives for Meta & TikTok @Bogzabs96]

Your customers have already written your best ad hooks.

You just haven't read enough reviews to find them.

After $107M+ in managed Meta spend, we built a Claude workflow that:

→ Pulls 500 reviews from your top sources (Trustpilot, Amazon, Reddit, surveys)
→ Clusters the language by pain, objection, transformation, and surprise
→ Turns customer quotes into 25-50 hook variants ready for Meta

Same system we run internally on every new Growthub account.

Comment 'REVIEWS' for the full 5-page workflow.

(Must be following)

[BuildTheDumb @probettinglab]

@ukhomeoffice Like Starmer tried to get rid of disappearing whatsApp messages, then used them himself with half the cabinet. Your a joke it wont work.

[Alex Northstar @NorthstarBrain]

Fun fact: your AI costs are going through the roof

But here are 10 ways to stop wasting tokens:

1. Start a new chat when the goal changes

Old context pollutes new work and consumes more tokens.

2. Give the AI a brief, not a brain dump
Goal. Constraints. Input. Output.

3. Upload only the files that matter
“Analyze my business” with 19 docs is how you reach your Claude limits.

4. Use summaries as handoffs
End long chats with: “Summarize the current state, decisions, open questions, and next steps.”

Then restart.

5. Use retrieval instead of dumping everything
Search first. Feed only relevant new chunks.

6. Trim tool outputs
Logs, JSON, HTML, transcripts, scrape results: clean them before sending.

7. Ask for deltas
“Only show what changed.”
“Only return the patch.”
“Only list the 5 highest-impact fixes.”

8. Cache repeated context
If your app sends the same system prompt, company docs, or policy text every time, cache it.

9. Make memory boring and short
Persistent instructions should be compact.
If your memory file becomes a novel, it will be expensive.

10. STOP always using the most expensive model!

Small model for simple tasks.
Big model for hard judgment.

In June I’m helping my clients learn the new AI skill: Context discipline.

building a guide for reducing AI token waste based on our best practices.

Comment or DM “TOKEN” and I’ll send it.

[Miss Ally @MissAlly_01]

Polce toay have sa they are nvestgatng a strng of ID thefts.

[MKingscott @M_Kingscott71]

@MayorKingAI There's this, too...

[MKingscott @M_Kingscott71]

New nightmare unlocked 😃