gnat35

@btcmos 沉稳且自带节奏，指的是ai coding都已经爆火了最近几个月才开始组建harness团队，到现在还没有产品吗😅

梁文峰和杨植麟的不同之处在于： 梁文峰， 通过幻方量化早都赚麻了。 杨植麟 还在基础创业的漩涡中挣扎。 所以 DeepSeek 给人的感觉就是沉稳且自带节奏。 Moonshot（Kimi K3） 给人的感觉还是有些浮躁和饥渴。 我从他们各自团队的人的讲话性格 就能感觉出来。

@MaxForAI 他们harness做的就挺晚的，到现在产品也还没发布

这人究竟是谁，竟能让梁文峰避他锋芒

@amritwt They are building their harness team right now, which is a couple of months late. That could explain a lot.

Deepseek fell off or something

@bojie_li 也正常

A rant about DeepSeek's interview process. First, after passing the written test, they didn't schedule an interview for half a month. By then, offers from other companies had already come through. Only after repeated prodding did they finally schedule the interview. When I interviewed at other companies, none of them required a coding interview, yet DeepSeek still required a written test, and the first two interview rounds were both coding interviews—something I never encountered while interviewing with any other company. During the second round, the interviewer felt that all the problems I described were research problems and had no grasp of what I was talking about; he kept pushing me to talk about what engineering challenges there were. During the coding interview, I had two monitors, and the interviewer said I kept glancing at the left screen, accused me of copying code, told me to stop copying, and said that if I couldn't prove I wasn't copying code, the interview couldn't continue. I felt deeply insulted and immediately decided to terminate the interview. Based on the conduct of both interviewers in the first and second rounds, DeepSeek turned out to be completely different from what I had imagined. I have decided to terminate DeepSeek's interview process, and I'm exposing it on social media. Please spread the word—feel free to share.

@ryanlpeterman @FeinbergVlad cool

Vlad Feinberg (@FeinbergVlad) is Google DeepMind’s pre-training area lead and I asked him all about how to land a job at a frontier lab like Google DeepMind, Anthropic or OpenAI. In this episode: • Skills frontier labs need • Differences between software engineering and AI research • Domains that matter for frontier research • Concrete steps engineers can take to get closer to research • Jeff Dean spot bonus story Where to watch: • YouTube - youtu.be/cDyi91onoJ8 • Spotify - open.spotify.com/episode/5XgGwE… • Apple Podcasts - podcasts.apple.com/us/podcast/the… • Transcript - developing.dev/p/google-deepm… Thank you to the sponsor of this episode for supporting my work: • WorkOS: makes your app Enterprise Ready with easy to use APIs to add SSO, SCIM, RBAC, and more in just a few lines of code, check them out at workos.com Chapters: 00:00 - Intro 00:33 - Skills frontier labs need 08:45 - The difference between AI research and engineering 21:41 - Domains that matter for the frontier 30:50 - Marketing yourself to frontier labs 35:13 - Concrete steps engineers can take 38:29 - Overview of pre-training areas 47:23 - Jeff Dean spot bonus story 50:14 - Favorite Gemini war story 58:59 - Advice for his younger self 01:03:07 - Outro

@devth @teej_dv @devinjameson you can try cursor cli

@teej_dv @devinjameson i want to play with composer but i ain't using their ide :/

Haven't used a Claude model for awhile. Used new 4.8 to test it out. Immediately started glazing me. Bro, just put the code in the bag

@ClaudeDevs genius idea: default fast😅

Fast mode now defaults to Opus 4.7 in Claude Code. Try it out today with /fast

hey @cursor_ai , for the past two days, I’ve been encountering multiple errors using the Cursor SDK while performing some daily tasks. model is Composer-2-Fast.

@ericzakariasson I want to see history chat using which model. It always auto switches to Composer, them I don't know whether i used GPT or Claude.

how can we make cursor 3 better? send us any bugs, feature requests, or feedback you have!

@leerob @ernielm our company use bitbucket, so local is a workaround

@ernielm Curious what use cases you have vs. cloud?

Cursor 3 keeps getting better! What should we add next?

@ThePrimeagen without energy, time means nothing

it's crazy how many people think that 30 minutes a day isn't enough to learn anything. Who told you this? Unknown knowns are wild. People just ingrain hurdles in their head about the universe because some thought leader spoke into their life in the distant past.

@NVIDIAGeForce GeForce Day

🟢 GEFORCE DAY IS BACK 🟢 To celebrate, we're giving away TWO GeForce RTX 5080 Founders Edition GPUs, signed by NVIDIA CEO Jensen Huang. Want one? Comment "GeForce Day" for a chance to WIN & stay tuned for more!

why can't openai add this kind web demo feature to gpt 5pro

@lexfridman @KeyuJin Kind of sad — you can only scratch the surface with these so-called “economists.”

Here's my conversation with @KeyuJin all about China. She is an economist specializing in China's economy, international macroeconomics, global trade imbalances, and financial policy. She is the author of The New China Playbook: Beyond Socialism and Capitalism. It's here on X in full and is up everywhere else (see comment). Timestamps: 0:00 - Introduction 0:47 - Misconceptions about China 5:17 - Education in China 14:34 - Economic reforms of Deng Xiaoping 19:53 - Mayor economy and GDP growth race 33:40 - Growing up in China 39:18 - First time in the US 43:32 - China's government vs business sector 47:06 - Communism and capitalism 50:45 - Jack Ma 56:58 - China's view on innovation and copying ideas 1:03:35 - DeepSeek moment 1:07:29 - CHIPS Act 1:09:16 - Tariffs and Trade 1:21:41 - Immigration 1:26:28 - Taiwan 1:32:14 - One-child policy 1:40:11 - China's economy collapse predictions 1:44:54 - Advice for visiting China

@ollama 699mb is kind of crazy.😅

Ollama's new app is here for macOS and Windows! Ollama 0.10! 🧵

@LeeStillfox framework + bundler

请教一下，在 2025 年这个 AI 时代。一个后端开发想要开始点前端技术栈，比较好的路线图应该是怎么样的？(在史前时代写过一点点的 bootstrap 和 jQuery)

@vikhyatk LLMs can certainly reduce the number of engineering jobs, but the sharpest tool is still just a tool. Some agent that calls different LLMs is still an LLM😅

> "in 2025, AI will replace mid-level software engineers" > reduces headcount by 5% not the actions of someone who actually believes what they're saying

@ThePrimeagen just put more time on fundamental stuff

its crazy how many doomers / e/acc there are in ai like.. just learn and become more and more useful and maybe one day those skills are less useful, but the hours and hours of problem solving and being creative probably will be useful for your whole life

@shengxj1 减肥需要改变长期的生活习惯，短期突击就算成功了不坚持也会反弹

无论是编程，房车，英语我行动力都还可以，从0学俩月php就出去面试，学俩月英语就海投，面试中学英语 只要开始行动，方法是否高效就可以，我们需要的不是【高效的学习方法】，而是【能实践下去的方法】 但是！在减肥方面，我发现我和群友一模一样，收集了大量的减肥方法，运动视频，资料，app一大堆 但是我从来都只开始减肥三天就放弃😭😭😭 没有行动

@sfnbz @shengxj1 因为Dan是网红，一般都知道他

@shengxj1 为什么说是Dan

我发现这个问题可以很好的鉴别是不是前端 你问js题，或者vue作者是谁，都可以用搜索引擎或者gpt造假 但是Dan这个错误答案 只有前端会回复这个 用搜索引擎和gpt只会得到Jordan Walke这个答案