开启时会模糊预览图，关闭后正常显示
🚨 Black Friday Alert 🚨 Today is the best day to invest in tools that save you time (and money). Here are my 4 biggest Black Friday deals👇 1️⃣ https://t.co/5hDIulxEEj — One extension to replace ALL your browser tools($29 Lifetime) 2️⃣ https://t.co/hwZ0eO0SVb — Search 300,000+ open-source SVG icons instantly($29 Lifetime) 3️⃣ https://t.co/PyEJHvyaGV — A self-hosted Mailchimp alternative you actually control($69 Lifetime) 4️⃣ https://t.co/RkKck3vLym — Beautiful invoices + API for developers($49 Lifetime) Buy now so I can stop refreshing Stripe every 5 minutes 😂💸
"The Micro Startups Guy" ❯ https://t.co/hwZ0eO0l5D ❯ https://t.co/RkKck3vdIO ❯ https://t.co/PyEJHvxCRn ❯ https://t.co/5hDIulx6OL Sold @nocodeapi for 6 figures
My new video. Laravel 3-Sided Marketplace: Structure Example (based on a real Upwork job) https://t.co/WScaotNb95
~20 yrs in web-dev, now mostly Laravel. My Laravel courses: https://t.co/HRUAJdMRZL My Youtube channel: https://t.co/qPQAkaov2F
RT @yan5xu: 看了眼 Anthropic 关于 Agent coding的新博客。只能说😂“早就想象到了”。 他们花大篇幅讲的“状态管理”和“上下文分片”机制，其实就是我之前一直在推的「双文档法」的规范版本。 当你发现自己的开发流（Workflow）不知不觉中契合了行…
Prompt Engineer, dedicated to learning and disseminating knowledge about AI, software engineering, and engineering management.
经常有学员做了产品，内测期间找我review，帮忙提意见，我都会认真的写文档给出一些意见。 会有我看着非常幼稚的作品，但是也会有一些惊艳的作品。比如今天学员的这个作品，完成度之高让我完全不相信是0基础的。但是真的才接触了2个多月！ AI时代真是太美妙了！人人都可以创作出自己的作品！
我的AI编程课(https://t.co/HVZn3ItASW) |B站up主 | 分享创造 + 无限迭代ing
to be fair - across providers, prompts and bench noise itself one has to spend a lot of time to do it properly. but if the reported scores have 10%+ gap from what you're reporting and you use it to claim to do better than others than burden of this additional effort is on you.
making models learn • eXperiments lab • memes and training lores
well if the baseline has such huge variance, it should be handled with multiple seeds and several experiments. mlebench has a good precedence for handling this where the performance is extremely noisy. 10%+ gap is quite a lot, i didn't bother to cross-check it yesterday, i should have.
to be fair - across providers, prompts and bench noise itself one has to spend a lot of time to do it properly. but if the reported scores have 10%+ gap from what you're reporting and you use it to claim to do better than others than burden of this additional effort is on you.