Inside Standard Bank’s multi-model AI bet
Standard Bank Group has named Amazon Bedrock as its platform and claimed a 20% gain from AI coding tools.
We haven't written up this one. TechCentral has the full story — the link below goes straight to it.
More in AI
How AI Is Making Mobile Accessibility Easier Than Ever
The OS is now covering for your mistakes: VoiceOver and TalkBack generate descriptions for unlabeled elements on device Automated audits catch more than they used to, but still not everything…
- AI enables automatic descriptions for unlabeled elements and images in mobile screen readers.
- AI used in development stages, from on-device assistive AI to AI-powered testing.
- Generative AI creates alt text, captions, and audio descriptions for mobile apps.
Stripe Bought the Routing Layer
Stripe just acquired OpenRouter, and everyone's talking about the number. Big deal. But the real signal is what layer they're buying into. OpenRouter is the router.
Models retire faster than operating systems
When an operating system deprecates an API , you get a year's notice, a migration guide, and a conference talk. When a model provider retires a model, you get a blog post, a date a few months out, and…
- Model providers retire models with minimal formal notice, unlike API deprecations.
- Models are catalog entries, not hardwired into code, allowing easy swapping.
- Retirement process involves catalog update, dialect adaptation, and evaluation suite testing.
AI’s trillion dollar token reckoning
We've now entered AI 2.0, where inference economics, data gravity, and control decide outcomes.
The Real AI Video Bottleneck Isn’t Generation. It’s Creative Control
AI video generation is getting easier. The next challenge is creative control: keeping characters, locations, story details, and visual style consistent.
Modern OCR narrows gap on doctors’ handwriting
A new benchmark of handwritten medical prescriptions shows that compact modern optical character recognition systems can decipher doctors’ handwriting far better than established OCR engines, while remaining well short of the accuracy needed for autonomous clinical use.
- GLM-OCR achieved highest Character Error Rate (CER) of 0.328 in April 2026 test
- PP-OCRv5 provided lower Word Error Rate (WER) of 0.789
- GLM-OCR correctly reproduced only 32.6% of tested words