Mike Griggs — creativebloke.com

Hi!

2026-09-28T00:00:00Z

I still do everything that I used to, Art and Creative Direction, CGI and Motion Graphics and VFX. However, I am fully embracing the fact that I have always liked how the things we use to make the pretty pictures work, as much as I like making the actual pretty pictures, hence the new job moniker.

This was born out of the past couple of years of deep diving how the latest batch of generative AI tools can create bespoke tools to make the creation of work easier, rather than just using it to make the images. I have never really fond using Generative AI to make images satisfying as it hides 'the journey' of creation and for both agencies and clients the journey is how we build trust in the creative process to enable the 'magic'.

So as well as being an Art and Technical Director with the nicest Creative Agency in the world StaticJoe Studio (opens in a new tab), I am available for opportunities to help people understand and create any stuff in the current world of 'WTAF is going on!'

My first After Effects Script is AE_CBrenamer

2026-09-30T00:00:00Z

My first After Effects script solves a problem I continually make for myself, which is that I am so eager to comp, pre-comp or make a comp from a Redshift render pass, that I can forget to name my comps as I go. This little script works in the After Effects Project Panel and gives numerous production-focused renaming options for Comps, Footage, or Solids.

I made a quick video to show how this script can be used.

AE_CBrenamer — on YouTubePlaying loads the video from YouTube, which sends your IP address and browser details to Google in the US. Their privacy policy.

I will be developing this as I go, so feedback is welcome, and the script is exclusively available on Sinilab.ai now. (opens in a new tab)

Cinema 4D MCP server announced

2026-09-30T00:00:00Z

I just downloaded the latest Cinema 4D release now which adds an MCP Server which can allow AI tools such as Claude to drive cinema 4D as can be seen from the screen grab.

I am very excited to see if I can use the new C4D MCP server in the same way that I have been using MCP in Blender in terms of creating assets and workflows that can benefit my work. Having MCP access in Blender has made software I 'kinda' know into an exceptionally powerful tool, that made me question how long I would need my Maxon One subscription for. I am hoping that the implementation of MCP in software I use every day (C4D) and know really well will be a game changer, or will it be something that gets in the way?

That being said, one of the things I do have concerns about with any creative software adopting AI workflows.

This is where I'm very grateful for my 30-year journey of banging creative tools with hammers to find out how they work. I have concerns that the journey that could be lost with an MCP workflow, and is that even relevant any more?

So many of our workflows that we've developed have been by making mistakes. And MCP software, especially with Blender, I found does make mistakes, but potentially that's in the prompting, not actually in the tools themselves. So that me understanding exactly what went wrong within Blender can be hidden from me by using MCP exclusively. The jury is out on whether that is good or bad, and artists like myself definitely need to reassess their value proposition. So it's all very exciting, but at the same time, I am coming at this fully open-eyed as to the challenges it could create.

Oh yeah, glossary time:

MCP = Model Context Protocol which is an open standard that acts as a universal connector for artificial intelligence tools, such as Claude for example, to securely talk to external tools, databases, and file systems which in this specific case is Cinema 4D.

View the launch video here :

C4D MCP Server — on YouTubePlaying loads the video from YouTube, which sends your IP address and browser details to Google in the US. Their privacy policy.

New Claude models? Test them before you switch

2026-10-08T00:00:00Z

Anthropic released Opus 5.5, Sonnet 5.5 and Haiku 5.5 this week for Claude. I usually just switch everything over on day one. This time, as I am trying to understand how agentic workflows can affect token use, I am happily going to follow my Scottish stereotype and I want to be 'tight' with my token use.

My creative bloke backend, that I use for a bunch of uses, including managing this site, has a spellchecking engine that I custom-made (I am especially proud of the 'DIMS' button, which checks to see if my usual garbled brain vomit is legible to another human). This is run through AI models, and I go for speed for the model choice.

Running multiple machines is great, but upkeep is a pain.

My house proofreader, the small service on my Mac Mini that checks my writing, asked for a model called “haiku”. That is a nickname, not a model. On my MacBook Pro it already meant Haiku 5.5. On the Mac Mini, which runs a slightly older Claude Code, it still meant Haiku 4.5. So the day the Mini updated itself, my proofreader would have changed model without anyone deciding it.

The fix is boring and important: in the script, name the model in full (claude-haiku-5-5), and have a test that refuses the nickname.

So I got Claude to run some tests.

Three tests, on real work

1. My internal proofreader: Haiku 4.5 against Haiku 5.5. The same planted mistakes and the same clean paragraphs, two runs each. Both caught 8 of 8 slips and invented nothing. Haiku 5.5 did it in 6.6 seconds against 17.1 = excellent and upgrade please to Haiku 5.5.

2. The search helpers: Haiku 5.5 against Sonnet 5.5. These are the small helpers Claude Code sends off to dig through my projects. I re-ran six real jobs from the last fortnight, word for word. Both got the main answer right on all six. Haiku was slower and read more, but each token costs a twentieth of Sonnet's, so a job came in at about a fourteenth of the cost. Haiku also found a real bug in one of my tools that Sonnet walked straight past. Thanks to the test, I am going to run with Haiku in the background, and hopefully I won't notice and also did I tell you I was Scottish.

3. The design critic: Sonnet 5.5 against Opus 5.5. I have a helper that gives an art director's crit of anything I am looking at (based on a lot of training of what I like when I am Art Directing, and this critical helper cuts a lot of wheat from the chaff and I can choose the number of options I want).

I gave both models four pages, one of them a copy of my proofreader page with six faults planted on purpose. Both found all six. Both gave the same pass or fail on every page, and named the same worst problem. Each caught a few things the other missed, which is a good argument for a second pair of eyes. So I need to make sure there is a degree of model flexibility on this process, which I will see if I can partially automate.

What actually saves tokens

What the test threw up though was that there is an even better way to economise on tokens. Every new message makes Claude re-read the whole conversation, and with the new million-token memory a chat could grow to nearly a million tokens before it tidied itself up. I now cap that at 300,000, and start a fresh chat with a short handoff note when the job changes.

So, from today:

Chats tidy themselves up at 300,000 tokens.
The search helpers run on Haiku 5.5.
The design critic runs on Sonnet 5.5, with Opus called in for a final look when it matters.
The main chat, where the planning and judgement happen, stays on Opus 5.5.

My biggest takeaway is never assume a new model is 'better' for you and your use case, (Gemini 3.1 Pro and Opus 5, I am looking at both of you when I say that). When a new model arrives test it against how you use A.I. as no one else uses it the way you do!

glossary:

Token = the unit an AI model reads and writes in, roughly three quarters of a word. It is what your plan or your bill is measured in.

Building an Linux AI testbed out of spares.

2026-10-14T00:00:00Z

One of the great things about my career so far has been the ability to accumulate kit. Whereas before I potentially would have sold it or left it unused and in a drawer, I'm finding that, with the combination of AI and a Tailscale network, I can repurpose kit in ways that could potentially make it useful.

I have a Mac Mini that I'm using primarily as a hub, which my AI tools can talk to and also write scripts for. Through the Tailscale network, that's become a useful server. The issue is that the M2 8 GB SOC on it is probably about to hit as much as it can do effectively.

I have this old Threadripper system that was becoming obsolete with Windows 10, and it had been a render box, but now I think I'm going to convert it to a Linux box to be an AI test bed with the potential to dual boot if I need a render node as well.

For the GPU, I'm using an AMD Radeon Pro W7800, which is going to enable efficient power use and has a fantastic 32 GB of RAM. I've been able to rescue 48 GB of RAM. Interestingly, I did have 64 GB of RAM, but one of the DIMMs had died, and it was causing issues. Thanks to help from my AI tools, I was able to diagnose which was the dead RAM chip, and now I've got 48 GB running in triple channel on this machine.

I haven't ever really used Linux before, but the incredible thing about working with AI tools is that, potentially, I don't need to know the core stuff, just what I require. I've already experimented in this workflow and integrated a Raspberry Pi4 into my system. Furthermore, I find the integration of AI tools to enable me to do things that I want without having to spend hours trying to find the right file. It is incredibly useful, and I'm hoping this is going to be the same for this system.

If you've got any questions about what I'm going to do, have comments about my messy workspace, or if you want me to document this process, please get in touch.

MCP on Cinema 4D is saving me £400 a year already

2026-10-23T00:00:00Z

Managing renders with Cinema 4D has always been a bit of a pain, which is annoying because it actually has some excellent features, such as Team Render Server. Team Render Server enables you to render on one device while still using Cinema 4D on another machine. If you need Redshift GPU, however, you do need a specific Redshift GPU render licence on every machine you are using.

Having a dropdown picture viewer on my MacBook Pro of what I am rendering on my workstation is comforting in all of the ways :)

For the past decade, I've been using Thinkbox Deadline to manage all of my local renders across multiple machines within my network. However, thanks to MCP access to Cinema 4D and the ability to use Cinema 4D's render preview, render queue, and team render server, I've been able to build a small render monitor.

I can talk with Claude to manage all of my render asks through multiple takes and multiple ways of rendering (such as the aforementioned team render server). I can also time renders to start as and when I want them to, and in whatever order I want them to. All of a sudden, MCP has enabled the equivalent of an incredibly sophisticated render manager for the cost of a Claude licence.

Claude can create reports on render for me with a single prompt

There was another catch with using a third party render solution. The issue was that, to use an external render manager such as Deadline (especially if you wanted to use Redshift across multiple machines with multiple Redshift GPU licences), Maxon insisted that you also needed an additional command-line renderer licence as well. That was an extra £400 on top of every single Redshift GPU render licence and a Maxon one licence.

MCP hasn't replaced the need for an external render farm such as Drop&Render in any way whatsoever. What it has done is enabled me to manage, through the Claude interface, all the boring stuff, such as take management, render settings, file naming and render file management.

Also, one of the key things about having this within my distributed system, which is managed across Tailscale, is that the Mac Mini also can send me emails about whatever I wish it to.

What I've done is I've tied in the render monitor to email me if there are any issues, because I might be out of the house, or working on something else. Having emails come in with alerts, I have found, as long as they're correctly formatted, is a great way to get notified about something that you can act on instantly, no matter where you are.

As I've said, MCP's killer feature for me with Cinema 4D (so far) is render management. I'm really interested to hear what your killer feature is?