Yuseferi

120 posts

Yuseferi

Yuseferi

@yuseferi

Katılım Nisan 2014
1.2K Takip Edilen379 Takipçiler
Yuseferi
Yuseferi@yuseferi·
@yahiyadev Never there will be Go v2. ;) it's against the biggest feature of golang which is backward compatibalibty
English
0
0
1
79
Yahiya
Yahiya@yahiyadev·
💀
Yahiya tweet media
QME
18
3
138
37.7K
Yuseferi retweetledi
Hugging Models
Hugging Models@HuggingModels·
How LLMs actually work: [📹 3Blue1Brown]
English
31
542
2.8K
164.6K
Yuseferi
Yuseferi@yuseferi·
@mitchellh depends, on how old it is, if it is less than three, still up key defeat `ctrl+r` in performance ;)
English
0
0
0
5
Mitchell Hashimoto
Mitchell Hashimoto@mitchellh·
Nobody should be using up arrows to get previous commands in a terminal or shell. You have to move your hand and its linear complexity in keystrokes. Use ctrl+p (for low n) or ctrl+r. Use a real shell or history manager (fish, fzf, atuin) for ctrl+r.
English
125
26
939
89K
Shikamaru ☕
Shikamaru ☕@pranavcodes_·
Be honest, can you write a 100 line code without using AI?
Shikamaru ☕ tweet media
English
533
17
884
58.8K
Yuseferi
Yuseferi@yuseferi·
I totaly agree with Bill! you shouldn't blindly use them, use should use them as a tool to make the work quality higher
William (Bill) Kennedy@goinggodotnet

I'm seeing an influx of posts about AI causing you to degrade your dev skills. I'm sorry, but that couldn't be farther from the truth. AI coding agents are a tool and you need to find the best tool for you and master how to use it. Would we tell anyone in any other profession they can't use the latest tooling to do their job better, smarter, more accurately, with some added performance? We've used tools like linters, vetting, and even to my dismay, do not edit code generation. Where were all these voices when we started using that tooling? Why is this tooling any different? It's not. I'm a better developer and engineer today because of this tooling. It's allowing me to move at a speed that was unheard of 2 years ago. My code is as good or better because I review everything and maintain my mental models. I'm learning better ways to do things, which in the past only came from mentors. As developers we have been asked to know way too much, and be experts in all of it. It's impossible. This AI tooling has moved that needle for me as well. Kronk has 18k LOC that I produced with help of AI over the past 3 months. When you count the FE it's close to 130k. Test me on any of the BE code and I'll tell you what it does, how, why, why it exists, and where it exists, both in the project and the architecture. I have a level of tests I've never had on any other project. This code base might be the most solid codebase I've ever produced. It's using everything I teach in my software design class. I'm not cutting corners, and I don't have to. It has cost me ~$200 in AI costs to build Kronk right now, and it's the best engineering money I have spent this year on myself. AI is not going to replace you, it's going to make you better. It's a tool, a really powerful tool. Learn how to use it effectively and stop trying to handcuff yourself and others.

English
0
0
0
29
Yuseferi
Yuseferi@yuseferi·
@alxfazio @mitchellh Ghossty is promising if it lets real engineers (not only medium-level AI agents) work on it, but to be honest, at the moment, it's overrated because of its creator! please open rooms for sound engineers to help and make it really good!
English
0
0
1
604
alex fazio
alex fazio@alxfazio·
i just ran vtebench with 16 panes (4 tabs × 4 splits), each running claude --dangerously-skip-permissions "please launch a 10 minute timer and monitor it as it goes", with the benchmark itself in an isolated 5th tab. this vaguely (very vaguley) resembles my actual workflow, but normally i have yazi, nvim and cc open across multiple tabs. turns out ghostty performs 2.21x better than kitty, at least according to the numbers. but in my actual day to day use, ghostty feels way worse. i get beachballs and slowdowns as soon as i have more than two tabs open. maybe cc is the problem, maybe yazi, maybe nvim, who knows. all i know is that between kitty, ghostty and wezterm, with the same workflow, ghostty is the only one that beachballs. kitty feels much better (yes it FEELS 10x better) despite being slower and doesn't get randomly stuck. i actually like ghostty and can't wait for it to improve, so if i can help you identify the problem, let me know
alex fazio tweet mediaalex fazio tweet media
English
4
0
76
13.8K
Yuseferi
Yuseferi@yuseferi·
@traits_reality But you can rely on a compiler that handle those complexity for you in a proper way ;). drink GREEN TEA with complier made it 40% faster and in another one you can wait more and more to get OWNERSHIP.
English
0
0
0
340
sin-ack
sin-ack@traits_reality·
you can ignore complexity, but you can't ignore the consequences of ignoring complexity
sin-ack tweet media
English
18
23
355
47.5K
Yuseferi
Yuseferi@yuseferi·
@goinggodotnet As always, sounds Great, Bill! 👏 I know you don't have much time, but spend more on open-source please
English
0
0
1
31
Yuseferi
Yuseferi@yuseferi·
@mitchellh and lots of "Remember Me" options that remember me as gold fish 🐟😀
English
0
0
0
23
Mitchell Hashimoto
Mitchell Hashimoto@mitchellh·
I just want to log in without being redirected 42 times or logged out every single day. I want to remain logged in on my device for at least months. We have machines that can mimic sentience and yet we can’t do log in for more than 24 hours. We’ve been played for fools.
English
105
82
2.1K
109.1K
Yuseferi
Yuseferi@yuseferi·
Just why #Golang is a better match with an agentic coder.
Yuseferi tweet media
English
0
0
0
33
Jeff Geerling
Jeff Geerling@geerlingguy·
I'm testing 1.5 TB of VRAM on a cluster of Mac Studios, courtesy of RDMA support in macOS 26.2, and @exolabs' new version of Exo (released under the OSS Apache 2.0 license today).
Jeff Geerling tweet media
English
77
156
2.1K
195.9K
Yuseferi retweetledi
Jaana Dogan ヤナ ドガン
Jaana Dogan ヤナ ドガン@rakyll·
Gemini 3 Flash is launching today. Frontier reasoning like a pro model, fast and efficient like a flash model. Gemini 3 Flash could be the most significant model launch from a quality/efficiency perspective to date.
Jaana Dogan ヤナ ドガン tweet media
English
3
7
117
9.5K
Yuseferi
Yuseferi@yuseferi·
@bardia_heydari I read this, it's very interesting! It gave me stronger inspiration of the path that I wanted to choose
English
0
0
1
1K
brdia 🇮🇷
brdia 🇮🇷@bardia_heydari·
کتاب با این شروع میشه که برنامه‌نویسا بعد ۵-۸ سال به سطح سنیور میرسن ولی اکثر جاها دیگه مسیر مشخصی برای رشد ندارن. اینجاست که خیلیا میرن سمت مدیریت ولی گزینه دیگه staff engineer شدنه. استف هارو بر اساس شخصیت و مسئولیت هایی که دارن میشه به چهار دسته تقصیم کرد:
brdia 🇮🇷@bardia_heydari

شروع کردم به خوندن این کتاب. براتون جالبه هر فصل خلاصشو اینجا بنویسم؟

فارسی
8
26
590
44.8K
Anton Zhiyanov
Anton Zhiyanov@ohmypy·
Accepted! Go 1.26 will include a SECRET package. Don't tell anyone about it 🤫 Just kidding. The package wipes the memory to keep secrets safe — hence the name. antonz.org/accepted/runti…
English
5
22
221
19.8K
Mitchell Hashimoto
Mitchell Hashimoto@mitchellh·
I’m very excited to share that my wife is pregnant with our 2nd kid! Just past the 1st tri and hoping for a healthy and successful pregnancy but really excited to grow our family and give my daughter a little brother (it’s a boy!).
English
178
9
3.6K
82K

Keşfet

@yahiyadev @mitchellh @pranavcodes_ @alxfazio @traits_reality @valyala @goinggodotnet @elonmusk