Keep on to blur preview images; turn off to show them clearly
昨天家人开了全局模式， 向 gpu 了服务器上传了数百个G 训练数据。 把我们家庭用的梯子这个月带宽都给用光了。 我一度以为断网了 ，然后怀疑是梯子不给力挂了， 就是没有想到会有这种情况。 然后我启用了racknerd 的作为备份服务， 太卡了。 我的自己体验： racknerd 作为梯子使用体验太差了，但是如作为早期项目服务器是可以的。
独立开发者 ｜ 个人IP教练 ｜ 帮助新手在X上完成早期成长｜ 公众号：PandaTalk8
To be clear - we are absolutely still in early iteration days, and I love to see new products shipped regardless of what they are 🫡 I’m just thinking about how to capture and migrate more consumer behavior over to AI native platforms - and there’s still work to do!
Partner @a16z and twin to @venturetwins | Investor in @gammaapp, @happyrobot_ai, @krea_ai, @tomaauto, @partiful, Salient, @scribenoteinc & more
做社交媒体、做SEO/GEO 、做产品， 都不是一个马上就能看到结果的事， 至少需要一年， 三年，五年。 但是这件事其实是确定的事， 因为你要只要肯定努力、肯坚持， 或多或少都能拿到结果。 只不过大多数人，今天做，明天就想看结果，后天就想发财。 抱着这样的期待，最后就是早早放弃。
独立开发者 ｜ 个人IP教练 ｜ 帮助新手在X上完成早期成长｜ 公众号：PandaTalk8
(this is sarcasm btw. For all I know they still have <10K of H100-equivalents. I do hope they've been building a new cluster, but there's no indication that it's played a role in anything up to V3.2. Their releases are consistent with 10K A100-40GB + 2K H800s + Ascend inference)
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1
can’t believe i used to schedule posts out to build a personal brand & post things that were just affirmations & quotes with no soul best strat is to just do cool things worth sharing and be curious u don’t need a personal brand, it’s a bad goal, better as a side effect
curious guy creating things @ https://t.co/HXWladhJaA - up and coming wife guy
There is this branch of the multiverse where I am an acclaimed movie director in Hollywood in the 50s. Probably a good friend of Ed Wood and Rod Serling.
You realize that I explicitly added the shoulder antennas?