MTS @cohere (prev: SrMLE @deliveroo, @heycurrent, ML for Music PhD @EdinburghUni, intern @Apple, ML MSc @UCL). Happiest when running/cycling/creating.
Edinburgh, UK
Joined August 2009
- Tweets2.9K
- Following1.1K
- Followers627
- Likes4.9K
Pinned Tweet
@weddingsketcher sketched my wedding. The results are absolutely gorgeous! This was a gift from my now wife @dianatbardsley.
Here they are (a thread!)
ALT A watercolour and pen sketch of people in suits and dresses outside the main entrance of a venue - the sign saying the biscuit factory is prominent.
James Owers-Bardsley retweeted
Who Gets to Define the Rules for AI?
Artificial intelligence needs evidenced standards, not a cartel. A perspective from @aidangomez, co-founder & CEO of Cohere:
cohere.com/blog/who-gets-to-…
Update - @LNER Ian at York managed to speak to the train guard and they have located the bag to hand it in at Bristol Temple Meads. What a hero, thank you Ian.
@CrossCountryUK I’ve left my grey eastpak suitcase by door in coach C on the 17:04 Edinburgh to Bristol CrossCountry service, which left York at 19:42. Could you please contact CrossCountry Control or the train manager urgently and have it taken off at Leeds and held for me?
If anyone around #Leeds #trainstation or #bristol #bristoltemplemeads @temple_meads can help I’d really appreciate it. Left suitcase in carriage C by door of the 17:04 from Edinburgh. Contains kid’s irreplaceable toy. All lost property lines advised have failed.
To update - a lovely @LNER person at York station has helped me. The head code of the train is 1V71. They’re trying to contact the guard to ensure the suitcase is safe.
@CrossCountryUK I’ve left my grey eastpak suitcase by door in coach C on the 17:04 Edinburgh to Bristol CrossCountry service, which left York at 19:42. Could you please contact CrossCountry Control or the train manager urgently and have it taken off at Leeds and held for me?
James Owers-Bardsley retweeted
Command A+ from @cohere is out now :) its our best model yet and its open source apache 2.0
James Owers-Bardsley retweeted
New @iclr_conf paper: "Learning is Forgetting: LLM Training As Lossy Compression" with @henryconklin Tom Hosking @yichern_tan Julian Gold, Jonathan Cohen, @cocosci_lab @max_nlp. Check out 🧵, blog, arxiv: henryconkl.in/posts/llms-are…, arxiv.org/abs/2604.07569v1 (entropy vid from blog)
James Owers-Bardsley retweeted
Today we’re launching Model Vault — a dedicated, fully managed platform to run Cohere models securely and at scale.
Model Vault delivers the control and isolation of self-hosting, without the operational burden:
🔒 Dedicated, isolated VPC
⚡ No noisy neighbors or rate limits
🔁 Elastic inference with predictable performance
📈 Real-time usage and performance monitoring
No sharing. No limits. No surprises.
Let Model Vault take care of model scaling and serving — so you can focus on building.
James Owers-Bardsley retweeted
How do we build multimodal systems that work effectively across the globe? 🌍
Today we release the Aya Vision Technical Report, the detailed recipe behind Aya Vision models, unifying state-of-the-art multilingual capabilities in multimodal and text tasks across 23 languages!
James Owers-Bardsley retweeted
I'm excited to the tech report for our @Cohere @CohereForAI Command A and Command R7B models. We highlight our novel approach to model training including the use of self-refinement algorithms and model merging techniques at scale. Command A is an efficient, agent-optimised multilingual model offering best-in-class capabilities. Read more below! ⬇️
James Owers-Bardsley retweeted
Today (two weeks after model launch 🔥) we're releasing a technical report of how we made Command A and R7B 🚀! It has detailed breakdowns of our training process, and evaluations per capability (tools, multilingual, code, reasoning, safety, enterprise, long context)🧵 1/3.
North is pretty great; my work contributes to the model behind it. I'm genuinely quite excited about avenues into healthcare and education which need this level of security, and am keen to see it tested.
Try it out!
James Owers-Bardsley retweeted
Introducing Command R7B: the smallest, fastest, and final model in our R series of enterprise-focused LLMs!
It delivers a powerful combination of state-of-the-art performance in its class and efficiency to lower the cost of building AI applications.
cohere.com/blog/command-r7b
James Owers-Bardsley retweeted
⌘-R
Introducing Command-R, a model focused on scalability, RAG, and Tool Use. We've also released the weights for research use, we hope they're useful to the community!
txt.cohere.com/command-r/
James Owers-Bardsley retweeted
First txt.cohere.com/command-r
