@paulg After many, many years of presuming good faith, I’m finding myself increasingly drawn to a Konmari curation policy: if the prospect of further interaction with a commentator does not spark joy, regardless of whether they are benign intent, block/mute and move on with life.
@syhily They fixed the max token reply thing in 0.9.0 so you can switch to DeepSeek using /model. Not very exciting, but /swarm is currently work in progress, so hopefully in a few weeks we’ll be able to use kimi to orchestrate and delegate to deepseek
@sheriyuo I think there's some miscommunication, you can just download from https://t.co/LV1VJPwpEW and replace your existing https://t.co/inG2Zyz9YY (assuming you're on macOS, idk about other platforms)
“From our perspective, I think the current direction of more cores, less power is the right way to go. At the end of the day, what we really value in an agentic world is the agent density, because CPU is currently the only thing that can materialize the intelligence that is produced by GPUs… The more cores with less power you can get without sacrificing the performance, the faster those intelligence can reach the real world and we want that to happen.” - Yujing (@gmi_cloud )
@zephyr_z9 I guess that’s why they’re withholding glm-5.1-highspeed from existing GLM coding plan subscribers to sell it to B2B customers for a better profit margin
@MikeFritzell If I’m not mistaken, that particular quick setup and collapse design was from a South Korean company, everyone held off for about six months before rushing in with copies 🥲
In January, Japanese SaaS stocks got crushed.
Claude Cowork made investors think that traditional software had been made obsolete.
Japanese fund managers were told not to touch anything software related.
New writeup on the other site (link in bio)
Focusing on Focus Point $FOCUSP: Malaysia's largest optical retail chain.
20% market share as the lowest cost franchise option amongst a fragmented industry.
Consistent low double digits grower at a 9x earnings multiple.