sbinnee

I wouldn’t have imagined cloudflare acquiring Deno. But AI got into both of them and somehow now it makes sense? It’s wild. Anyhow congrats to Deno team.

I am on the same boat. 3.5 medium was disappointing, glm 5.2 and 5.3 were very welcome. But I still cannot leave anthropic or openai. This chonky model being expensive, mistral’s subscription quota will run out ridiculously fast

I wouldn’t mind this monster to be the mascot for mistral 4 large. It’s cute

Yes, it arrived months late. The waiting was too long I started losing faith. But when I saw Mistral’s series funding news, I knew they are on the right track. 3 billion EUR is just too significant

It is a good approach to promote it as a German model. But I wonder if it really thinks like the German think. I speculate they just translated texts in other languages, likely English and Chinese, into German.

Congrats! I am surprised that you first chose Korea for PhD. As myself being Korean, I hear a lot of horrendous stories about PhD students in Korea. They tend to spend for more than +8 years. I know a few people who went through their PhDs and certainly it took +8 years for all of them. I did my PhD by the way in France and it took me 4 years.

Did you know that it could take so long when you first started your PhD in Korea? Could this be one of your reasons to move to Japan?

I love the form factor. I would love to tinker with it!

I cannot say that I agree to all the issues in the post. But men I am sick of meat proxies too.

> available on iPhone Duo, iPhone Air, iPhone 16 models or later, iPhone 15 Pro, iPhone 15 Pro Max

Such a high bar for iOS. Unfortunately new Siri is not for me.

Claude loved pandas, at least around 4.5 opus. I had to keep reminding it to use polars instead. I wonder if it’s changed since.

My raspberry pi 4 struggles every time I scan the library. I hope that this release helps.

I am still waiting for their large model. Seemingly they only recently started combining reasoning, agentic, and multimodal in one model. They have mistral-3.5 medium and mistral-4 small. There is no 3.5 small nor large not to mention no 4 medium nor 4 large. These model numbers to me seem like that their model release did not meet the schedule. I am still waiting for 4 large and expect it to match at least glm 5.2 performance.

Yet it is an impressive number. But yeah when you see a number 99 you have doubts. Thanks for the link

I dropped my claude subscription a few months ago, though I kept some credits to do this and that with claude, thinking that claude might do better for some tasks. A few days ago they were all expired. It feels like it’s time to let claude go.

I remember mistral provides some free credits for opting in data for training. And for some reason I could not find this option anywhere in my console. Does someone know if this plan still exists?

This is huge!

It is probably true that it has a big attack surface. But omarchy is no doubt a huge driving force. A few days ago I listened to a podcast dhh talking about the latest major release and its huge donation. I believe it’s now 10m usd or something. I am hopeful that dhh is serious enough to address the security issues plus a lot of ux improvement on linux.

In the episode he emphasized 17 layers of security layers where I remember the number solely because I found hard to believe to be honest.

You can find the podcast episode in this one https://thestanduppod.com/

It is not going to be cheaper though. I may choose the cheaper one in the end because performance will be marginal, both being flash.

While living in France, I finally learned how to hold my bladder for half a day like an adult

I used to run xorg and wayland together and switch between them because you know you setup xorg no matter what even if you daily drive wayland. Since 2024 or so however I don’t remember opening a xorg session.

I clicked this one with no hype. I am now exited to update the kernel of my raspberry pi 4.

I have been using openrouter only superficially. I didn’t know about the broacasting feature. Thanks for that.

Yes it does look like a good solution. But when I imagine actually using a guardrail for a product, this model only outputs yes/no probabilities. There is no reasoning trace why it was rejected. Users or even developers would have no idea why a prompt was classified yes or no. I really like this release but I feel like I need something more to use it as a guardrail in production.

I am involuntarily in a mini retirement. The very previous job turned out to be a disaster. I have been enjoying this opportunity by traveling a lot with my partner, exercising a lot more, meeting old friends, visiting family, playing unfinished games, and reading novels a lot. It’s been great. But there is always a certain pressure back in my mind for my career forward. I wonder if it’s become more or less common to hop between jobs every two to three years. I cannot see myself staying in a company for a decade long. Maybe it’s just because I was always working for startups, rather intense environments.

I wonder if it’s going to be some custom encoder/decoder using DNN. I would also want to see some examples!

I wholeheartedly agree on readme. AI generated readme irks me off almost immediately.

Openrouter also features routing. Routing is indeed an option if you don’t need consistent behavior and allow switching models.

If it offers more than opencode go, the entry plan looks enticing

I saw that metal is almost x10 faster than vulkan? Why so much gap?

Same here