rubslopes

Yes, they did, which is unfortunate, as the game is very good in itself. They gave a nice twist to the formula.

Why? Price? If the reason is performance, I've been using non-frontier models for cheap, and they run great for my needs (GLM 5.2, DeepSeek v4 Pro).

Wow. Sci-fi stuff!

Same here, but I live in a developing country. I usually switch to international news (or tech news, or no news at all) because local news is mostly about murders and other terrible things. I have family in Portugal, and when I visited, I remember the local news being mostly about pastries and folklore events. I watched TV every day while I was there.

I've been using Zen since almost the beginning of the project. I'm pretty happy with it! Great QOL features, and being Firefox, I have perfect ad block and syncs nicely with Firefox for Android.

I was going back and forth between kimi k2.6 and deepseek v4 pro, but it's been 2 weeks that I've been using deepseek alone. It's very good and very cheap, I pay the opencode go $10 subscription and don't have to worry about quotas.

I've met some of the IT team of the city government of Rio de Janeiro years ago, before LLMs were a thing. They were already doing state-of-the-art data science projects. Glad to see how far they've gone.

If it's really fully compatible with docx files, I'm very interested in this. No matter the advances Apple makes with the M chips, Office apps are always slow to open. Not to mention the obligatory background services.

Samsung routines can do this.

MCP allows you to easily separate API requests from their access tokens, so that the LLM only has access to the requests part. Giving an LLM CLI access removes all boundaries, anything goes.

EDIT: to add an example: I have a personal claw agent that I only use CLI, I don't care. But I'm also building an agent inside a company product, and there we use MCP all the way.

I feel lucky to have been promoted to a management position recently, just as I was starting to feel less excited about dev work because of AI. I still enjoy building systems, but I have to admit that the loss of challenge made the work much less enjoyable for me.

Now I have a team of interns to mentor. They're sharp and use AI constantly, so my guidance is less about code and more about UI/UX, understanding what the client actually wants, good work practices, well-documented tickets, thorough reviews, and so on. Thankfully, I like this work, it has been very rewarding.

Yes, I do think it is.

HTML made by Claude will, by default, be "sleek, modern", with colorful tables, cards, maybe Tailwind for styling. And, of course it will, if you wanted a barebones HTML, you would just have asked for markdown!

So the LLM decided to present some content using 4 cards, and you now want to add new itens. You can't just add new lines of text: you need to copy the whole HTML of the cards. But the LLM used different colors for each card, so now you have the first cards with varying colors and the new cards all the same color as the last card. Now you have to think about colors... etc etc

Although it will make up for a "wow, we really did it" moment, it will be met with hollowness

I hadn't heard this view before, but it does make sense...

People are afraid to touch dangerous things, like passwords and payment systems. Depending on their skill level, they should indeed be afraid.

Correct.

Still, the day we manage to run a full LLM on biological neurons, even if using conventional code under the hood, will be a very interesting day for consciousness discussions.

I thought that that was the case for me, but then I tried using Claude Code through the desktop app last week and it was so bad. Slow, glitchy... I went back to the TUI in no time.

I don't think the real-world evidence supports your argument... OpenAI and Anthropic have all of those advantages today, and Chinese models are reaching the same level. Clearly, the Chinese labs are doing something very right that is not directly related to infinite money.

Happy birthday to the youngest and oldest!

I think OP means standardized benchmarks.

A fellow user replied below, but it refers to the software that uses the LLM (Claude Code, Opencode, pi.dev, etc.).

---

Funny you mention that, because I started noticing the word 'harness' being used everywhere about a month ago, even though I hadn’t seen it before (in this context). As I don’t trust my memory, I assumed I had just been overlooking it and added it to my vocabulary. However, a Google Trends search does show increased usage since the end of March: https://trends.google.com.br/trends/explore?date=today%203-m...

Harness: opencode

Subscription: opencode go

I also use a claw agent[1] via Telegram, which uses pi.dev under the hood with my opencode go subscription.

[1] I forked one of those Claw projects (bareclaw) and made many changes to it.

April has been a crazy month for open weights models. I've been using Claude Code for work and Kimi 2.6 for personal projects and Kimi has been very good. Glm-5.1 is also great. Qwen, Mimo and Deepseek I need to test some more, but they all have been producing good results. I have the impression that they are all are at the same level, or close to, Sonnet 4.6.

Thanks for the tip. Just tried it and it worked great between MacOS and Android.

90% here, and it makes sense. I'm very picky about saying something is blue!

I was confused about why your comment was being downvoted; it sounded like an honest opinion... Until I got to the last sentence. You wrote a self-fulfilling prophecy.

If this was true, they should have contacted the user nevertheless.

Was this thread invaded by AI? Casually reading the first comments, 3 different users mentioned they had a recent "mental health incident" related to caffeine??

Search this page for "mental health incident"

I had more luck finding "similar minds" in academia.

Before GLM-5.1, I was going back and forth between Opus 4.5 and Kimi 4.5 and having very good results with Kimi.

My first child is still 16 months old, and I already feel like this.