Full Claude Course
Master Claude to automate repetitive tasks, build apps without coding, create real portfolio projects, and become more valuable at work.
في عام 2015 حدث في كوبا مشهد يبدو وكأنه خرج من فيلم رعب .
رجل يدعى بابلو موريرا كان مصورا من هافانا ويبلغ من العمر 28 عاما بحسب التقارير التي تناولت قصته قال بابلو إنه كان يمر بمرض شديد جعله يفقد القدرة على الاحتفاظ بالطعام ويخسر وزنه كما وصف وجود مشكلات واضطرابات كبيرة في حياته المنزلية رحل للاطباء ولم ليعرفوا الحل
ثم اقترح عليه أحد معارفه زيارة معالج روحي يعتقد أنه يستطيع إزالة ما وصفه بابلو لاحقا بأنه لعنة فودو وهنا يبدأ الجزء الذي جعل الفيديو
ينتشر حول العالم جلس بابلو أمام المعالج ووضع المعالج بيضة في فمه ثم بدأ ما يشبه طقسا روحيا وبعد لحظات ظهر وكأنه يستخرج من فمه أشياء غريبة جدا
ووضعها داخل وعاء بابلو قال لاحقا إنه كان متوترا لدرجة أنه لم يدرك ما يحدث أثناء الطقس وأنه لم ير ما تم استخراجه إلا بعد أن نظر إلى الوعاء
وبحسب روايته فقد اعتقد أن تلك الأشياء كانت جزءا من اللعنة التي أصابته
لكن هنا يجب أن نفصل بين ما قاله بابلو وبين ما يستطيع العلم إثباته
وجود الفيديو حقيقي كفيديو متداول ووجود بابلو موري��ا وقصة تصوير المشهد في مقاطعة ماتانزاس عام 2015 ورد في عدة تقارير
لكن لا يوجد دليل طبي او علمي موثوق لان بكل بساطة العلم لا يعرف بالغيبيات وانما بالادلة
Kimi K3 is going to slaughter OpenAI and Anthropic for enterprise and here’s why:
Yesterday I had a post that went pretty viral thanks to a retweet by @ClementDelangue (big thanks 🙏🏼)
It was about how I was downloading the K3 weights.
A lot of the comments were asking what I was going to do with them.
And the answer is, I’m going to be running tests for a law firm on the cloud to see if we can hit internal benchmarks with K3.
Because right now, the firm spends nearly $30k/mo in API tokens on Claude opus and fable.
But if we can achieve similar performance with K3, the firm is committed to spending around $500,000~ on hardware to run this locally.
We’re looking at a Gigabyte G893-ZX1-AAX4 which comes with 8x AMD MI355X which gives you about 2.3TB of VRAM and has been validated on Kimi (about $400k)
But here’s the thing, they won’t be buying it outright, they would actually finance it over five years (pretty common term on hardware)
And that’s what makes this very interesting.
Because if they finance this over five years they’re looking at a payment around $10,000~ on the hardware and probably like $18k/mo if you add in everything like electricity and colocation.
And by doing this, they will unlock:
- more token usage
- local data privacy
- unrestricted access
- a savings of 20-30% on their monthly AI payment
And it gets even better for US companies
Because Trump passed the one big beautiful bill which allows US businesses to deduct 100% of their hardware and GPU purchases in year zero
So not only would the firm be able to get more usage, more privacy, more control, and a lower payment, they would also get to deduct 100% of the purchase the same year even while being financed 🤯
I seriously don’t see how OpenAI and Anthropic maintain $1T valuations anymore
Because the moment a company hits about $20k in monthly API costs, it starts to make sense to consider financing your own hardware.
Wild times ahead.
Open source is clearly a winner. 💸
*none of this is financial advice*
Se você é um homem que tem:
30 Anos
32 anos
33 anos
34 anos
36 anos
37 anos
35 anos
E está reconstruindo sua vida, Comece com essas 12 coisas.
1. Conserte seus dentes, pele e postura
@FotosDeFatos Esse filme tem a mesma duração que esse aí @FotosDeFatos mas passa uma lição de vida que muitas jovens sofrem no dia a dia
Vc sabe qual é?