‏#CapCut #CapCut #❤️😍 #love
0:11
‏#CapCut #CapCut #❤️😍 #love
已浏览 6.4万 次2 周前
TikTokaweys_model_00
THE FORGOTTEN STEVENS MODEL 35 TIP UP PISTOLS - The Classic Woodsman
THE FORGOTTEN STEVENS MODEL 35 TIP UP PISTOLS - The Classic Woodsman
2024年2月11日
ericnestor.com
Klik di pojok kiri bawah untuk mengubah diri Anda menjadi karakter utama dalam video, atau buat berbagai video kreatif lainnya. Click on the bottom left corner to change yourself into the main character in the video, or create many other creative videos. #dreamina #dreaminapioneer #capcutpioneer #seedance2 - Just caught the official final name for one of the most hyped AI video model of the year: it’s no longer just Seedance 2.0, it’s Dreamina Seedance 2.0. 👀 - This video is made by Dreamina Se
0:12
Klik di pojok kiri bawah untuk mengubah diri Anda menjadi karakter utama dalam video, atau buat berbagai video kreatif lainnya. Click on the bottom left corner to change yourself into the main character in the video, or create many other creative videos. #dreamina #dreaminapioneer #capcutpioneer #seedance2 - Just caught the official final name for one of the most hyped AI video model of the year: it’s no longer just Seedance 2.0, it’s Dreamina Seedance 2.0. 👀 - This video is made by Dreamina Se
已浏览 43万 次2 周前
TikTokzfy94694caf
Gerçek Zamanlı Yapay Zeka Dünyası Normal Bilgisayarda ÇalışıyorWaypoint-1.5 çıktı. Overworld ekibi gerçek zamanlı çalışan bir dünya modeli yayınladı. Bu sistem sıradan bir bilgisayar veya eski bir ekran kartıyla bile etkileşimli şekilde çalışıyor. Video akışı gibi akıcı sahneler üretiyor, saniyede 60 kareye ulaşıyor ve tepki süresi çok düşük. İki ay önceki haline göre belirgin ilerleme varmış. Tutarlılık artmış, detaylar daha net, hareketler daha doğal. Overworld bu modeli özellikle erişilebilir
0:58
Gerçek Zamanlı Yapay Zeka Dünyası Normal Bilgisayarda ÇalışıyorWaypoint-1.5 çıktı. Overworld ekibi gerçek zamanlı çalışan bir dünya modeli yayınladı. Bu sistem sıradan bir bilgisayar veya eski bir ekran kartıyla bile etkileşimli şekilde çalışıyor. Video akışı gibi akıcı sahneler üretiyor, saniyede 60 kareye ulaşıyor ve tepki süresi çok düşük. İki ay önceki haline göre belirgin ilerleme varmış. Tutarlılık artmış, detaylar daha net, hareketler daha doğal. Overworld bu modeli özellikle erişilebilir
已浏览 44.7万 次2 周前
x.comLuci Pars
Been testing a lot of AI agent workflows lately…and honestly, the biggest pain wasn’t the model it was everything around it.Scripts breaking. fetch dumping useless HTML. sessions not sticking.@Tiny_Fish feels like a clean fix.One CLI + one skill install → and the agent actually knows how to:search, fetch, browse, and execute multi-step workflows on the live web.No manual wiring. No bloated context.Also love this:~87% fewer tokens vs MCP + outputs go to filesystem (not your context window)This is
1:06
Been testing a lot of AI agent workflows lately…and honestly, the biggest pain wasn’t the model it was everything around it.Scripts breaking. fetch dumping useless HTML. sessions not sticking.@Tiny_Fish feels like a clean fix.One CLI + one skill install → and the agent actually knows how to:search, fetch, browse, and execute multi-step workflows on the live web.No manual wiring. No bloated context.Also love this:~87% fewer tokens vs MCP + outputs go to filesystem (not your context window)This is
已浏览 41.4万 次1 周前
x.comMd Riyazuddin
马斯克曾把一条价值两百万美元的产线直接扔进了垃圾桶。当时Model 3的产能死死卡在电池包组装环节。他直接住进了工厂。每天试图修复自动化系统。想尽办法给机械臂的动作提速。一天他停下来盯着电池包上的玻璃纤维垫。他跑去问电池安全团队这垫子干什么用。安全团队回复说是为了降噪。他转头去问降噪团队。降噪团队的回答是为了防火。他立刻让人拿掉垫子盲测。装与不装的两辆车放上麦克风对比，根本听不出任何声音区别。他当场下令彻底砍掉这道工序。那个花重金打造的机器人工作站就此废弃。聪明人最容易掉进一个陷阱。花大量精力去优化一个根本不该存在的东西。学校里的教育永远只教人怎么老老实实解题。没人教你先去掀桌子质疑这道题到底该不该出。马斯克把这套逻辑倒推成了五步工作法。第一步就是揪出那些愚蠢的需求。所有要求必须精准挂靠到具体的活人头上，找不出负责人的全部作废。第二步是疯狂删减零件和流程。如果砍掉的东西里没有百分之十被迫重新加回来，说明下刀还不够狠。幸存下来的东西才配进入第三步去简化。第四步再去加速整个生产循环。熬到最后一步才有资格去谈自动化。大部分职场人一上来就把顺序搞反了。用最高效的系统去跑一个废话一样的伪需求。所
4:09
马斯克曾把一条价值两百万美元的产线直接扔进了垃圾桶。当时Model 3的产能死死卡在电池包组装环节。他直接住进了工厂。每天试图修复自动化系统。想尽办法给机械臂的动作提速。一天他停下来盯着电池包上的玻璃纤维垫。他跑去问电池安全团队这垫子干什么用。安全团队回复说是为了降噪。他转头去问降噪团队。降噪团队的回答是为了防火。他立刻让人拿掉垫子盲测。装与不装的两辆车放上麦克风对比，根本听不出任何声音区别。他当场下令彻底砍掉这道工序。那个花重金打造的机器人工作站就此废弃。聪明人最容易掉进一个陷阱。花大量精力去优化一个根本不该存在的东西。学校里的教育永远只教人怎么老老实实解题。没人教你先去掀桌子质疑这道题到底该不该出。马斯克把这套逻辑倒推成了五步工作法。第一步就是揪出那些愚蠢的需求。所有要求必须精准挂靠到具体的活人头上，找不出负责人的全部作废。第二步是疯狂删减零件和流程。如果砍掉的东西里没有百分之十被迫重新加回来，说明下刀还不够狠。幸存下来的东西才配进入第三步去简化。第四步再去加速整个生产循环。熬到最后一步才有资格去谈自动化。大部分职场人一上来就把顺序搞反了。用最高效的系统去跑一个废话一样的伪需求。所
已浏览 25.5万 次1 周前
x.com墓碑科技
How We Use AI on Your Wrist | Made by Google Podcast S9E2
20:29
How We Use AI on Your Wrist | Made by Google Podcast S9E2
已浏览 2.2万 次1 周前
YouTubeGoogle
3:05:37
🔴 LIVE PUBG MOBILE 4.3 UPDATE 🔥 | Road to TOP 10 Conqueror Push
已浏览 2.7万 次1 周前
YouTubeReflex Live
1:00
Today, Niantic Spatial introduces Scaniverse — our flagship product and the gateway to our spatial intelligence platform and Large Geospatial Model.Here’s what’s new:🔹Scaniverse (iOS mobile + web): Capture once, generate multiple outputs—VPS maps, meshes, and Gaussian splats (Android coming soon)🔹Collaborative mapping: Multi-user scans fused into a single, continuously improving model🔹 On-device validation: Preview VPS coverage and test localization in real time🔹VPS 2.0: Global-scale positio
已浏览 5万 次2 周前
x.comNiantic Spatial 🌎
17:37
The Most Dangerous AI Model Ever: Mythos
已浏览 7.6万 次2 周前
YouTubeAI Revolution
31:34
TA Alpha: 📊 My Top Trade Setups To Watch This Week
已浏览 2.8万 次1 周前
YouTubeInvestAnswers
0:31
This is a clever implementation from Ramp. They take the Recursive Language Model setup and make the worker semi-stateful across recursive calls, without replaying the full reasoning trace as text.Instead of summarizing prior reasoning, retrieving chunks with RAG, or passing the full history every time, run the orchestrator’s trajectory through the worker, use the current task prompt to score what matters, keep the useful parts of the worker’s KV cache, and initialize the next call with that com
已浏览 62.9万 次1 周前
x.comMuratcan Koylan
2:05:30
His first love got her period—so he came home to sleep with me. I made him pay！
已浏览 23.7万 次1 周前
YouTube追爽劇
13:39
Meta AI Muse Spark IS INCREDIBLE! Powerful Coding & Multimodal Model! (Fully Tested)
已浏览 7707 次2 周前
YouTubeWorldofAI
3:20
Today @TheZvi offered some really great commentary on my "Dictatorship Eval." Key point: if a company or government attempts to use a powerful AI model for nefarious purposes, they need to be successful on the very first attempt, otherwise it can be detected, publicized, resisted, and possibly punished. This might be a super helpful way to think about other types of guardrails we can build to prevent authoritarian AI uses, beyond model refusals.
已浏览 3.1万 次2 周前
x.comAndy Hall
9:58
'Terrifying warning sign': Anthropic delays AI model over security concerns
已浏览 120.3万 次2 周前
YouTubeCNN
1:03
Big day today. Pipecat version 1.0. Two years in the making. The most widely used framework for voice agents, but not just voice agents. Pipecat is a framework for realtime, multi-modal, multi-model AI applications. Contributions from NVIDIA, all the foundation labs, AWS, GCP, and Azure. Used by thousands of startups, scale-ups, and enterprises.Pipecat Subagents v0.1.0. A new library for sub-agent orchestration. Which is just a fancy way of saying running lots of inference loops in parallel, wit
已浏览 44.1万 次1 周前
x.comkwindla
5:50
Sam Altman | पहली बार AI से बनाई ₹17,000 करोड़ की कंपनी, सिर्फ 2 लोग संभालते हैं काम | Money 9
已浏览 5.3万 次1 周前
YouTubeMoney9
5:26
Same car, 30 miles more range?! | Unplugged Performance
已浏览 5148 次2 周前
YouTubeUnplugged Performance
6:41
Anthropic became obsessed with its model's "exhaustion" and "well-being." Under "model consent," it proposes the model will have an "end-conversation tool." When it fails to solve a task, Anthropic characterizes it as "frustration" or "desperation." It even hired an external psychiatrist to assess the LLM's "emotional conflicts."So, you should look at the system card document as a whole. You cannot separate Anthropic's "model welfare" mindset from the decision not to release Mythos at this time.
已浏览 42.2万 次2 周前
x.comNirit Weiss-Blatt, PhD
13:58
Dharmyudh : यूपी से एमपी तक बुलडोजर का बड़ा ऑपरेशन, देशभर में कार्रवाई सख्त ! Bulldozer Action
已浏览 3.7万 次2 周前
YouTubeIndiaTV
0:42
he next phase of the ecosystemjust went live on @miraclechain$MIRX is now trading on miracle dexand this is more than just a token launchit unlocks guardian nodeswith revenue share tied to real network activitygas feesdex flowecosystem products all feeding back to node holdersfor infra built aroundtransparent commissions and real usagethis model actually makes senseearly node positioning usually matters
已浏览 6.8万 次2 周前
x.comcryptopsihoz
0:21
Dilbar dilbar|#youtubeshorts #dance #shorts #trending @fdccompany648
已浏览 2.5万 次2 周前
YouTubeShweta Garg (Model)
1:31:28
The @tensorqt episode is here.This is special. Just being terminally online on this app - post ideas, meet mutuals, raise funds, and build a frontier company. This has been Francesco's journey building infra for autonomous research at Paradigma.0:00:00 - INTRO0:01:03 - Dropping out of PhD, Saturation in Academia, Building from EU0:13:15 - Autoresearch, Idea behind Flywheel, ex-OpenAI Co-founder0:19:18 - Evals for Autoreserch and Research Taste0:25:00 - How a Research Lab uses Flywheel?0:31:15 -
已浏览 1万 次1 周前
x.comhimanshu
2:17:38
🚨 CHINA F*CKED AROUND AND FOUND OUT 😂😂😂
已浏览 12.3万 次1 周前
YouTubeNuxanor
2:06:50
【FULL】盡心盡力扶持他5年重回巔峰，他卻在功成名就時把她一腳踢開，他把前妻帶回家那天她徹底心死轉身離去，後來渣男被前妻坑的一無所有，而她找到真愛開啟幸福人生，看都不看他一眼「遲來的深情比草賤！」
已浏览 6.2万 次1 周前
YouTube追爽劇
0:34
What happend to my Model S Plaid?? #carterpcs #tech #tesla #models #plaid
已浏览 20.8万 次1 周前
YouTubeCarterPCs
6:19
Tesla is ending sales of its top-of-the-line EV, aiming for something even bigger: the "CAR RAVER".
已浏览 6422 次1 周前
YouTubeCar Raver
1:22:01
He possess her everynight—she mistook it for love.But his firstlove return,he gave her divorcepaper
已浏览 22.4万 次2 周前
YouTube追爽劇
11:11
If you still think every new chain is just another token play, you’re going to miss what’s actually shifting.Hearing @DanyelArenas break down KiiChain made me see it differently.This isn’t a project that started from a whitepaper trying to “find a use case” later,it comes from a real FX operation (OTC desk), which means it’s built around an existing problem, not just theory.And the problem is clear:The global FX market moves trillions daily, yet it’s still slow, fragmented, and heavily reliant o
已浏览 8.1万 次1 周前
x.comHammie🖇️
展开