Interesting! If they have this, why are the Chinese labs still so compute-limited (hence relying on distillation instead of training from scratch), and why do they seem to want Nvidia chips so badly?
Everyone, including the Chinese labs, seem to understand that Huawei's chips are not as capable as their press releases make them out to be.
Also, for what it's worth, tech-insider.org is an embarrassingly lazy AI-generated content farm: https://tech-insider.org/about/
I would hazard a guess that Huawei is paying to get positive stories about their chips on the English-language internet, because they certainly aren't getting any organically!
They are not as capable certainly but considering they didn't have anything like 3-4 years ago, I'd say the potential to catch up is there.
Also like OpenAI's Jalapeno, with LLMs it might be feasible to deploy relatively simpler silicon that's hard to program for, and farm out implementation to AI so the CUDA moat might be shallower than before.
>I would hazard a guess that Huawei is paying to get positive stories about their chips on the English-language internet, because they certainly aren't getting any organically
Jensen has said that 7nm Huawei chips are comparable to A100-level chips, and even fairly recent models as of 1>year ago were trained on those and process improvements are not what they used to be
Just spreadsheets for the time being. I think irrespective of the tool it's more the mindset for me. What is too small to have meticulously documented versus a general bucket. Also what becomes too big to not track on a ticket anymore but via project governance.
I didn't say French was weird; I said the messages are weird, and they happen to be in French.
I don't speak French, and using auto-translate on a few, they seem to be innocuous, context-free replies to message chains I was never in. I think the goal is just to get me to respond in some way to start the chain, like those spam text messages that just say something like "Great, Ted, Wednesday works." and if you respond, they're off to the races working to establish a relationship with you they can cash in on.
1. I use both Melious[1] and Entrim[2]. I have a Melious Pro subscription, €150/month and use Entrim pay-as-you-go API, because it's cheap.
I used to have two Mistral subscriptions, because I needed the extra capacity, but quickly realised I'd need three or more, so switched to the two providers above.
2. Providers covered in the first point. Regarding models I use a combination of GLM5.3, GLM5.3 Flash and Deepseek 4.1 Flash 1017. Which model I use depends on the task. GLM5.3 Flash for general chat and co-ordination tasks. Deepseek 1017 for most coding related work and GML5.3 for adversarial batteries and documentation.
3. Maki[3], although quite customised with various plugins. No python permitted, only Janet for scripting. Extremely limited bash calls, instead preferring to use the built in harness tooling, or plugin tooling I've written. AGENTS and skills used extensively. Mechanical gates also used to prevent certain actions, like remote git pushes without explicit prior approval.
I switched to Maki for the token saving strategies and stayed for the ergonomics, which I think I like, but maybe just feel comfortable with now.
4. None listed. The closest I come is a multipane Zellij setup with different agents running in different panes. Orchestration is still human in the loop though.
Deepseek 4.1 Flash 1017 has been pretty good for coding, I agree. It doesn't seem to do documentation as well, things like requirements, specifications, implementation plans. Here I still use GLM5.3 for a better outcome. I also use both models as adversarial batteries to check the work of each other.
Having said that, 1017 has been great in saving money. It's available generally much cheaper than GLM5.3, and I think it delivers >90% of the capability.
> Ends in 'a' = female: a relatively safe assumption, as far as an assumption here can be safe, with no other context. A strong majority in many places will match that.
Which places? I guess I ask because your next sentence:
> Many of the exceptions will be obviously foreign names with traditional spelling
Which makes me wonder if you meant "US places", because I can think of a few male names, which maybe categorised as "foreign" if you're in the US, that end in "a".
e.g. Csaba (pronounced "Chubba") is a reasonably common Hungarian name.
And some not so "foreign" ones, albeit mostly with Biblical roots, like Joshua and Ezra.
The names with Biblical roots did not end in -a in their original form, but in most cases their termination was "-ah". The final consonant was usually dropped in English, resulting in the "-a" termination.
reply