https://xitter.cf/t.co/dOepAobNK8 helps people to speak using only their eyes! Author of Blazor-University, Fluxor, etc Coding for over 40 years. Happily married to @lyndor

Joined February 2025
What do you think of this tattoo our daughter did? I think it's bloody brilliant!
8
231
Note: If you need to contact me in a crypto secure fashion, please encrypt your message with the following private key... 4
3
1
11
572
Anyone else in/near Birmingham UK going to this AI coding event tonight? meetup.com/producttank-birmi…
4
157
I watched Reform Exposed on channel 4. At Reform's request, a team paid £30k for polls that appeared in national papers. Reform didn't declare this as a donation, which is illegal. The team were journalists. I want to know who put up the £30k! Was it a donor for another party, therefore it was indirectly an illegal political donation to that party? @Channel4 @c4dis
46
Calling people inbred because they have Pakistani origins. Tommy "I'm not racist" Robinson. Not racist at all, noooo.
1
1
131
A friend is working on Chameleon, a Material 3 Expressive component library for Blazor, React and Web/JS. Angular (Vue and Svelte support is planned by Feb 1st, 2027). chameleonui.io/blog-entries/…
3
140
Hey @YouTube - why am I shadow banned? My comment isn't showing up. Here is my channel - youtube.com/channel/UCzqFMoT… Here is my comment youtube.com/watch?v=6c8fpShF…
2
2
295
Those who fail to learn from history are doomed to automate it.
6
272
This is the kind of thing I want to see AI doing!
Replying to @CERN
@CERN and @MistralAI are building the first foundation model for collider physics. It's gonna be trained on billions of simulated LHC collisions, one exabyte of real data 🤯 This will help us accelerate the processing of data and by extend, our understanding of the universe! Love to see this happening in Europe!
3
269
I wonder if a decision model like Layla could be used in Claude/OpenCode/Codex etc to decide based on the prompt which tools' descriptions to include in an AI request context? @opencode
2
178
AI thinking levels explained.
2
165
I watched a video about Local AI on a plane. Verdict: Much better than snakes!
3
133
Currently downloading an AI model at 48 MB/s. My first owned modem (back in the 90's) used to take approximately 9 minutes per MB. First I borrowed (1200baud) would have taken nearly 2 hours per MB.
3
140
Downloading now 👍
Run DeepSeek V4.1 Flash 1017 on 2x DGX Sparks ✨ - 1M total KV pool, 512K context - 75 tok/s on prose, single stream - 130 tok/s on prose, 4 concurrent streams - Prefill 2000-2175 tok/s - EXL3-EAQv1, 2.9 bpw The quant in this recipe got about the same calibration as my GLM 5.3 Flash, with some extras. Feedback would be appreciated! Special thanks to @bertholomusai and all the contributors for the engine work, and to @ashxhart for TensorFold. Get it here: github.com/MiaAI-Lab/DeepSee…
1
9
581
Peter Morris retweeted
I asked Opus 5.5 to refactor some C# code. I wanted some derived bool properties moved from UI into the domain class. It negated the expressions on ALL of the 5 or so properties, which broke the app. You can't trust AI.
2
1
2
100
My Dutch Barn advert idea. I wonder if JK Rowling would object? #DutchBarn @rickygervais
82
Bah, 512GB still not available
4
9
434
This is what GPT claims for M3 Ultra vs 2xGB10 running GLM 5.3 Flash. (M5 estimated to follow)
5
1
2
1,107
If this is accurate, it seems it's not worth me spending money on an M5 Ultra is it? Possibly not even to link up to my 2 GB10s via MCDMA.
1
126