Daniel at Reading Advantage

Reading Advantage เปิดโปรแกม blended learning สำหรับชั้นประถมศึกษาแล้วครับ ต่ามสเปคของ Cambridge YLEครับ reading-advantage.com/th/services/bl…

@paradite_ Muting abusive people? Yes. Muting people you just disagree with? Why, unless you don't trust your own beliefs enough to have them challenged.

I’m muting people on X who don’t understand how far ahead Anthropic is. It’s hurting my brain so much that I had to do this. If you think OpenAI is in anyway better than Anthropic, please just block me to save both us some trouble.

This is the way. I was going to say that the obvious answer is git and rsync. Sysadmin.

Mimo pro is an excellent model on par with Sonnet 4.6. the multimodal (standard) version is slightly less capable, but still great, and better for daily tasks.

@XFreeze @business Everyone I know uses Doubao.

@business most Chinese people I've met don't know anything about ChatGPT, Grok or Gemini or Claude They said they only use DeepSeek

DeepSeek’s delayed release of its V4 model points to a strategic shift toward deeper integration with China’s domestic chip ecosystem, according to a social media account affiliated with the government-controlled CCTV bloomberg.com/news/articles/…

@dev_maims I think you're right in one way In my day, we did assembly and C, gone-- juniors who cared to learn white using AI will absolutely dominate, but too many are just using it to chat and not learn fundamentals, which will keep them from fine detail work.

Unpopular opinion: junior devs who learned with AI will outperform seniors in 2 years.

@svpino I'm currently running five in parallel on one project, doing an audit. There are 21 games to audit, five at a time, with one subagent on each.. It's been going for four hours and it's only at game 15. I also have Cron jobs that fire off throughout the day.

I don’t believe people who say they are running “12 parallel coding agents”. Either they are lying for clicks, or I’m a complete retard who can barely keep up with a single Claude instance.

Five Kimi subagents running on OpenCode in parallel at a time, one per game. For four hours. I'm at 61% of my 5-hour limit. I'm not even sure what I'd have to do to max out the usage.

I think I may have overbought my @Kimi_Moonshot coding plan with Allegretto. I have a repo of 21 mini games I've developed for my devs to import into Reading Advantage. I set Kimi on an audit of 29 shared contracts all games must have. 1/2 continued.

@Samaytwt Because it's much easier to install Linux on.

Give me one solid reason.. to choose a Windows laptop over a MacBook

Gpt-5.5 really is quite amazing. * I have a monorepo of next.js apps using convex that I migrated in from separate projects. * Problem starting dev server. * Mimo can't do it. Then GLM-5.1 can't do it. Then GPT-5.3 high can't. * Finally, 5.5 medium cracked it. Congrats.

@burkov Acquihire in addition to users. X.ai lost a lot of founding engineers once their stock options matured.

Cursor is Elon's first purchase, which is a huge mistake. A coding agent harness is now open source (see Codex and Claude Code). The current design works virtually perfectly, so there's no need for a fundamentally new design. One can say that agentic coding is solved. One might argue that he bought Cursor for users, but users, as we know, switch between coding IDEs frictionlessly, so it's not like Twitter, where you cannot leave without losing followers. Cursor is an emperor with no clothes. A shell without substance. You only know it exists because it's been there before most others.

@LewisCTech The answer is simply network effects. Unless someone develops a disruptive vision for what LI does, it's not going to be displaced.

Why is there no competitor to LinkedIn? - It's slow - It's full of AI slop - It actively encourages people to write AI slop - It's slow - The timeline is awful - It's slow There is clearly a market for a professional social network. I would jump in a second.

Google just released the DESIGN.md spec. This is an easy way to do a site redesign and create a unique look for each of your projects. github.com/google-labs-co…

@llmdevguy @MiniMax_AI I bought the max plan of minimax for a year, and get virtually unlimited use, even images, videos, music, and tts. M2.7 is definitely not a great coding model, but it's ok-ish if you define the work into small chunks and have a better model come behind to review every few phases.

GLM 5 -> 5.1 - Great improvement Kimi K2.5 -> K2.6 - Great improvement DeepSeek V3 -> V4 - Great improvement Qwen 3.5 -> 3.6 - Great improvement I am looking at you @MiniMax_AI. When 3.0? 😀

@kimmonismus This looks just like my Chinese "Ark Coding" plan (volcengine). Is it the same but different name for China and global?

@Gamingtronium Our time is valuable. Installing Linux Mint takes almost no time and works out of the box. But that's never the real pain of setup. It's making it work FOR YOU. Macs take a lot longer on that step for me. Linux takes almost no time for the things I do.

Linux is only free if your time has no value 🤷🏻‍♂️

@rsuyoy Current MacBook or a then current MacBook? Give it a few years, and computer specs will increase to allow this.

What's a reasonable timeline for a GPT 5.4/Opus 4.6 level LLM that runs locally on a good Macbook 10 years? much less? much more?

PISA math is also reading. reading-advantage.com/th/blog/the-pi… #education #readingadvantage

@petersteele My problem is that I have only $20 subs for Claude and Codex, so i also have subs with Minimax, Kimi, Bytedance, and Xiaomi. Altogether maybe $140, but basically never run out of inference this way.

What the hell are yall doing that you need to spend so much on AI for? I literally have 1 sub, paid through work, for 100 a year to GH Copilot. I literally go into each new billing cycle with less than 70% used. Right now, its end of April and I am not even at 50%. I can't fathom multiple hundreds a month in AI subs and still needing more, like WTF are yall doing?