Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

Two models. Zero rivals.​ Total dominance. DBX707 or DBX S. DBX. No Equals.​ Discover: #AstonMartin #DBX #NoEquals #POWERDRIVEN

astonmartin's profile picture

Aston Martin

1,528,648 subscribers

15,177 görüntüleme • 9 ay önce •via X (Twitter)

Bilim & TeknolojiOtomobil & Araçlar#AstonMartin#DBX#NoEquals#POWERDRIVEN
LIVE

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

Class leading-performance.​ Accept nothing less.​ DBX707 or DBX S. DBX. No Equals.​ Discover: ​ ​#AstonMartin​ #DBX​ #NoEquals​ #POWERDRIVEN
0:32

Class leading-performance.​ Accept nothing less.​ DBX707 or DBX S. DBX. No Equals.​ Discover: ​ ​#AstonMartin​ #DBX​ #NoEquals​ #POWERDRIVEN

Aston Martin

10,958 görüntüleme • 9 ay önce

Luxury mastered in every demanding detail.​ Benchmark set. Again. And Again.​ DBX707 or DBX S. DBX. No Equals.​ Discover: ​​ #AstonMartin​ #DBX​ #NoEquals​ #POWERDRIVEN
0:35

Luxury mastered in every demanding detail.​ Benchmark set. Again. And Again.​ DBX707 or DBX S. DBX. No Equals.​ Discover: ​​ #AstonMartin​ #DBX​ #NoEquals​ #POWERDRIVEN

Aston Martin

11,066 görüntüleme • 8 ay önce

"(DBX S) is a brilliant car" - The Telegraph Discover DBX S: #AstonMartin #DBXS #POWERDRIVEN
0:30

"(DBX S) is a brilliant car" - The Telegraph Discover DBX S: #AstonMartin #DBXS #POWERDRIVEN

Aston Martin

13,181 görüntüleme • 10 ay önce

We’ve always loved a challenge, Bentley Motors. Discover DBX S: #AstonMartin #DBXS #POWERDRIVEN
1:09

We’ve always loved a challenge, Bentley Motors. Discover DBX S: #AstonMartin #DBXS #POWERDRIVEN

Aston Martin

13,284 görüntüleme • 1 yıl önce

An SUV doing non SUV things. The fastest, most powerful in its class. Discover the record-breaking DBX S: #AstonMartin #DBXS #POWERDRIVEN
0:58

An SUV doing non SUV things. The fastest, most powerful in its class. Discover the record-breaking DBX S: #AstonMartin #DBXS #POWERDRIVEN

Aston Martin

15,070 görüntüleme • 11 ay önce

In Parliament today I proposed optional Joint filing of Income Tax Returns for married couples. Family A Both spouses earn ₹10 lakh each. Total household income is ₹20 lakh. Tax: zero Family B One spouse earns ₹20 lakh. The other stays home to raise their child. Total household income is ₹20 lakh. Tax: ₹1.92 lakh The only difference is how the salary is split between the two spouses. One roof. One kitchen. One household budget. But when tax time comes, the family disappears. The tax system sees two individuals. A husband and wife become strangers. No clubbing of income or rebates. In Parliament today I proposed optional joint filing of Income Tax Returns for married couples, so families with uneven incomes are not unfairly penalised. If implemented, then Family A and Family B both will pay Zero Tax.
4:36

In Parliament today I proposed optional Joint filing of Income Tax Returns for married couples. Family A Both spouses earn ₹10 lakh each. Total household income is ₹20 lakh. Tax: zero Family B One spouse earns ₹20 lakh. The other stays home to raise their child. Total household income is ₹20 lakh. Tax: ₹1.92 lakh The only difference is how the salary is split between the two spouses. One roof. One kitchen. One household budget. But when tax time comes, the family disappears. The tax system sees two individuals. A husband and wife become strangers. No clubbing of income or rebates. In Parliament today I proposed optional joint filing of Income Tax Returns for married couples, so families with uneven incomes are not unfairly penalised. If implemented, then Family A and Family B both will pay Zero Tax.

Raghav Chadha

1,112,067 görüntüleme • 4 ay önce

🚨 TRUMP ESCALATES: MORE BOMBING TONIGHT ON IRAN, BIGGER & MORE POWERFUL President Trump is done playing games. After Iran downed a U.S. Apache helicopter near the Strait of Hormuz, America has already struck two nights in a row. Now Trump is warning another round is coming tonight, and he’s eyeing Kharg Island, Iran’s critical oil hub. From Truth Social: The U.S. will hit their Navy, Air Force, radar, anti-aircraft, and take total control of their oil and gas infrastructure, just like we did successfully in Venezuela. No more drip-drip delays. No more letting Iran run out the clock on nuclear weapons. Maximum pressure is working. The regime is bracing. This is what strength looks like. Peace through overwhelming power. Trump is putting America First and forcing the deal, or total dominance.
6:17

🚨 TRUMP ESCALATES: MORE BOMBING TONIGHT ON IRAN, BIGGER & MORE POWERFUL President Trump is done playing games. After Iran downed a U.S. Apache helicopter near the Strait of Hormuz, America has already struck two nights in a row. Now Trump is warning another round is coming tonight, and he’s eyeing Kharg Island, Iran’s critical oil hub. From Truth Social: The U.S. will hit their Navy, Air Force, radar, anti-aircraft, and take total control of their oil and gas infrastructure, just like we did successfully in Venezuela. No more drip-drip delays. No more letting Iran run out the clock on nuclear weapons. Maximum pressure is working. The regime is bracing. This is what strength looks like. Peace through overwhelming power. Trump is putting America First and forcing the deal, or total dominance.

Gunther Eagleman™

20,551 görüntüleme • 2 ay önce

Meet the F-15: America’s Undisputed Air Superiority King ✈️💥 In a world where speed decides survival, one legend rises above them all. The F-15 Eagle — not just a fighter… but America’s undisputed master of the heavens. From the moment its twin engines ignite, it doesn’t take off — it erupts. Mach 2.5. Thrust that defies gravity. A design so pure it was built for one thing only: total air dominance. No compromises. No equals. Watch it scream off carrier decks, carve through the clouds like a blade, and own every inch of the sky. Over 100 confirmed air-to-air victories. Zero losses. A record written in fire across deserts, oceans, and decades of conflict. This isn’t just metal and missiles. This is American engineering at its fiercest — the eagle that never blinks, never yields, and never loses. From the runway to the stratosphere… the king doesn’t ask for permission to rule. It simply takes it. F-15 Eagle — Undefeated. Unmatched. American. 🇺🇸
0:59

Meet the F-15: America’s Undisputed Air Superiority King ✈️💥 In a world where speed decides survival, one legend rises above them all. The F-15 Eagle — not just a fighter… but America’s undisputed master of the heavens. From the moment its twin engines ignite, it doesn’t take off — it erupts. Mach 2.5. Thrust that defies gravity. A design so pure it was built for one thing only: total air dominance. No compromises. No equals. Watch it scream off carrier decks, carve through the clouds like a blade, and own every inch of the sky. Over 100 confirmed air-to-air victories. Zero losses. A record written in fire across deserts, oceans, and decades of conflict. This isn’t just metal and missiles. This is American engineering at its fiercest — the eagle that never blinks, never yields, and never loses. From the runway to the stratosphere… the king doesn’t ask for permission to rule. It simply takes it. F-15 Eagle — Undefeated. Unmatched. American. 🇺🇸

NancyH

99,611 görüntüleme • 3 ay önce

🚨 do you understand what this AMD engineer just proved on camera watch the clip, it’s all there in one frame: llama.cpp slot logs on the left, nvitop showing 79% GPU util holding steady, localhost chat answering a 26k token prompt at 711 tok/s prompt eval, the gemma-4-12B Q4_K_XL gguf running clean. no cloud call, no api key, no censorship layer. for 2 years the rule was: frontier models need the cloud, local is too slow for real work, pay $200/month and wait for tokens. today a $1,800 lunch box runs qwen 3 235B at 11 tok/s with zero data leaving the room, and the same box runs gemma 4 12B with multi token prediction at 14+ tok/s on 30k context with no throttling. $440/month subscription stack: dead. $200/month claude tier: dead. local is too slow excuse: dead. one lunch box. $9/month in electricity. two moats buried in one week. the you need the cloud line wasn’t protecting users it was protecting the subscription.
0:21

🚨 do you understand what this AMD engineer just proved on camera watch the clip, it’s all there in one frame: llama.cpp slot logs on the left, nvitop showing 79% GPU util holding steady, localhost chat answering a 26k token prompt at 711 tok/s prompt eval, the gemma-4-12B Q4_K_XL gguf running clean. no cloud call, no api key, no censorship layer. for 2 years the rule was: frontier models need the cloud, local is too slow for real work, pay $200/month and wait for tokens. today a $1,800 lunch box runs qwen 3 235B at 11 tok/s with zero data leaving the room, and the same box runs gemma 4 12B with multi token prediction at 14+ tok/s on 30k context with no throttling. $440/month subscription stack: dead. $200/month claude tier: dead. local is too slow excuse: dead. one lunch box. $9/month in electricity. two moats buried in one week. the you need the cloud line wasn’t protecting users it was protecting the subscription.

AdiiX

22,223 görüntüleme • 1 ay önce

FREE DeepSeek V4 Pro and V4 Flash API. Runtime by Bad Theory Labs just dropped it. Same platform, new free model, same 2-minute setup. > Runtime already gave you free Claude Opus 4.8 and GPT 5.5 access. Now they added DeepSeek V4 Pro and V4 Flash to the free tier. Absolutely FREE, YES. Same exact flow as before: → Go to runtime(.)badtheorylabs(.)com → Sign up with Google → Fill in your details during onboarding → Free balance lands automatically(But you don't need it to use DeepSeek - it's completely free right now) → Grab your API key from the dashboard → Plug into OpenCode, OpenClaw, or Hermes - done DeepSeek V4 Pro is one of the strongest coding and reasoning models available right now. Free tier also includes btl-2 - auto-routes between GLM, OpenAI, Anthropic, and Gemma. 10 million tokens per month at zero cost. 344 models total. One OpenAI-compatible endpoint. No code rebuild, no card, no limits hunting. >> Just start using the API, and when choosing a provider, select DeepSeek I tested it myself - fully working. Bookmark this before the free tier fills up.
0:23

FREE DeepSeek V4 Pro and V4 Flash API. Runtime by Bad Theory Labs just dropped it. Same platform, new free model, same 2-minute setup. > Runtime already gave you free Claude Opus 4.8 and GPT 5.5 access. Now they added DeepSeek V4 Pro and V4 Flash to the free tier. Absolutely FREE, YES. Same exact flow as before: → Go to runtime(.)badtheorylabs(.)com → Sign up with Google → Fill in your details during onboarding → Free balance lands automatically(But you don't need it to use DeepSeek - it's completely free right now) → Grab your API key from the dashboard → Plug into OpenCode, OpenClaw, or Hermes - done DeepSeek V4 Pro is one of the strongest coding and reasoning models available right now. Free tier also includes btl-2 - auto-routes between GLM, OpenAI, Anthropic, and Gemma. 10 million tokens per month at zero cost. 344 models total. One OpenAI-compatible endpoint. No code rebuild, no card, no limits hunting. >> Just start using the API, and when choosing a provider, select DeepSeek I tested it myself - fully working. Bookmark this before the free tier fills up.

Atenov int.

26,920 görüntüleme • 1 ay önce

i pointed hermes agent at nvidia's nemotron cascade 2 30B-A3B on a single RTX 3090 24GB. IQ4_XS quant by bartowski, 187 tok/s, 625K context. had it discover its own hardware, create an identity file, then build a full GPU marketplace UI from a single prompt. it one shotted it. first attempt no iteration. qwen 3.5 35B-A3B on the same hardware same 3090 24GB took an iteration to recover from a blank screen on the same type of build. 24 days between these two models releasing. same active parameters, completely different architectures and cascade 2 through hermes agent just keeps going. this model goes on and on. feast your eyes. more iterations and tests dropping soon. nvidia really cooked. no special flags needed. nvidia optimized this mamba MoE so well it just runs. flash attention auto enabled, context auto allocated. the model does the work not the config. but i compiled llama.cpp from source and i'm not sure how it performs on other engines. if you ran nemotron on any hardware drop your numbers below. RTX, AMD, Mac, whatever. model, quant, tok/s, engine. i want to see if it holds everywhere or just on llama.cpp.
1:31

i pointed hermes agent at nvidia's nemotron cascade 2 30B-A3B on a single RTX 3090 24GB. IQ4_XS quant by bartowski, 187 tok/s, 625K context. had it discover its own hardware, create an identity file, then build a full GPU marketplace UI from a single prompt. it one shotted it. first attempt no iteration. qwen 3.5 35B-A3B on the same hardware same 3090 24GB took an iteration to recover from a blank screen on the same type of build. 24 days between these two models releasing. same active parameters, completely different architectures and cascade 2 through hermes agent just keeps going. this model goes on and on. feast your eyes. more iterations and tests dropping soon. nvidia really cooked. no special flags needed. nvidia optimized this mamba MoE so well it just runs. flash attention auto enabled, context auto allocated. the model does the work not the config. but i compiled llama.cpp from source and i'm not sure how it performs on other engines. if you ran nemotron on any hardware drop your numbers below. RTX, AMD, Mac, whatever. model, quant, tok/s, engine. i want to see if it holds everywhere or just on llama.cpp.

Sudo su

70,981 görüntüleme • 4 ay önce

MLX + TurboQuant = Local Super Power Take a local (private) document(s) or codebase, pre-fill the 256k KV cache (the context) with the document(s) and system prompt, quantise and run on Apple's MLX and you have almost instantaneous, lossless document queries with total privacy. For a 75-page PDF (some 30k tokens), I was seeing sub-150ms TTFT (Time To First Token), that's "instant" for a human. We run this on an office server (one of our Mac Minis) for staff to work with, instant answers on critical, confidential company/client documents on phone, iPads or laptops. Imagine the "What were the 3Q25 profits?", "When did Joe start working for us?", "Please list the project milestones and team count" answered within seconds of hitting the return key. It works with voice too. This example is 17k lines of code from 177 source files plus an index of the entire 1,902 files, the entire Claude Code codebase (oops Anthropic!). Zero loss (unlike Anthropic's security). What this means is the ability to work better on local (private) code bases, off-the-shelf private document analysis with better response than the best public models - all locally.
0:33

MLX + TurboQuant = Local Super Power Take a local (private) document(s) or codebase, pre-fill the 256k KV cache (the context) with the document(s) and system prompt, quantise and run on Apple's MLX and you have almost instantaneous, lossless document queries with total privacy. For a 75-page PDF (some 30k tokens), I was seeing sub-150ms TTFT (Time To First Token), that's "instant" for a human. We run this on an office server (one of our Mac Minis) for staff to work with, instant answers on critical, confidential company/client documents on phone, iPads or laptops. Imagine the "What were the 3Q25 profits?", "When did Joe start working for us?", "Please list the project milestones and team count" answered within seconds of hitting the return key. It works with voice too. This example is 17k lines of code from 177 source files plus an index of the entire 1,902 files, the entire Claude Code codebase (oops Anthropic!). Zero loss (unlike Anthropic's security). What this means is the ability to work better on local (private) code bases, off-the-shelf private document analysis with better response than the best public models - all locally.

John T Davies 🇪🇺

85,589 görüntüleme • 4 ay önce

Eric Schmidt (former Google CEO) laid out an AI timeline at the Future Investment Initiative in Riyadh in October 2024: Within 5 years (2029–2030), AI systems will start writing and recursively improving their own code—triggering an explosive upward slope in capabilities. Then, 6–8 years from now (2030–2032), a single non-human system could reach 80–90% of top-expert performance across every field simultaneously: physics, chemistry, art, medicine, and beyond. No single human (not even Leonardo da Vinci) could match that breadth. Schmidt: "We don't know what happens when such a thing exists... but we know that that race is really important." This matches his consistent warnings through 2025 (e.g., recursive self-improvement in 4–5 years, AGI-level dominance soon)—and with 2026's rapid breakthroughs, the clock feels even shorter. A single AI smarter than any human in everything is coming fast. What scares or excites you more about this timeline—utopia, existential risk, or total job transformation?
1:03

Eric Schmidt (former Google CEO) laid out an AI timeline at the Future Investment Initiative in Riyadh in October 2024: Within 5 years (2029–2030), AI systems will start writing and recursively improving their own code—triggering an explosive upward slope in capabilities. Then, 6–8 years from now (2030–2032), a single non-human system could reach 80–90% of top-expert performance across every field simultaneously: physics, chemistry, art, medicine, and beyond. No single human (not even Leonardo da Vinci) could match that breadth. Schmidt: "We don't know what happens when such a thing exists... but we know that that race is really important." This matches his consistent warnings through 2025 (e.g., recursive self-improvement in 4–5 years, AGI-level dominance soon)—and with 2026's rapid breakthroughs, the clock feels even shorter. A single AI smarter than any human in everything is coming fast. What scares or excites you more about this timeline—utopia, existential risk, or total job transformation?

Camus

15,554 görüntüleme • 5 ay önce

PC Andrew Harper was dragged to his death at 42mph. His uniform was torn off and his injuries were so bad that first responders thought he was a deer carcass. The men who killed him showed zero remorse they even laughed in court. They got just 13 years each for manslaughter. Now, under Labour’s new early release scheme, two of them could walk free after only 6.5 years. This is a total insult to PC Harper, his family, and every police officer. Killing a cop and showing no regret should mean real time behind bars — not a slap on the wrist because prisons are full. How can any government release such evil monsters back onto this country’s streets especially after they have shown no remorse or empathy of their atrocious crimes. Any government that is prepared to release such evil into society is itself an enemy of the people.
1:15

PC Andrew Harper was dragged to his death at 42mph. His uniform was torn off and his injuries were so bad that first responders thought he was a deer carcass. The men who killed him showed zero remorse they even laughed in court. They got just 13 years each for manslaughter. Now, under Labour’s new early release scheme, two of them could walk free after only 6.5 years. This is a total insult to PC Harper, his family, and every police officer. Killing a cop and showing no regret should mean real time behind bars — not a slap on the wrist because prisons are full. How can any government release such evil monsters back onto this country’s streets especially after they have shown no remorse or empathy of their atrocious crimes. Any government that is prepared to release such evil into society is itself an enemy of the people.

Benonwine

3,147,859 görüntüleme • 20 gün önce

Greg Osuri built the case for decentralized AI in three layers in under 5 minutes. Layer one: access. Anthropic's Claude is the best coding model available. OpenClaw is the most successful open-source project in history. Anthropic just announced they're banning OpenClaw usage. The most powerful model cutting off the most loved tool. The company that controls the best intelligence can revoke your access whenever their interests change. Layer two: transparency. You have no idea what biases Anthropic has built into Claude. If AI models directly impact how you think, the decision tree needs to be visible. Closed-source models with hidden biases making decisions that shape your worldview isn't a feature, it's a risk. Layer three: compute. Amazon may be picking winners in AI by controlling compute allocation. Cloud companies stopped being neutral marketplaces and started giving GPUs to the companies they've bet on. Amazon overwhelmingly backing Anthropic might have contributed to Claude's dominance. The cloud is no longer a utility. It's a kingmaker. Greg Osuri 🇺🇸's conclusion: intelligence should be free like water and air. Decentralized AI isn't ideology. It's the only architecture where the model can't ban your tools, the biases can be audited, and the compute isn't allocated by a company picking favorites.
4:33

Greg Osuri built the case for decentralized AI in three layers in under 5 minutes. Layer one: access. Anthropic's Claude is the best coding model available. OpenClaw is the most successful open-source project in history. Anthropic just announced they're banning OpenClaw usage. The most powerful model cutting off the most loved tool. The company that controls the best intelligence can revoke your access whenever their interests change. Layer two: transparency. You have no idea what biases Anthropic has built into Claude. If AI models directly impact how you think, the decision tree needs to be visible. Closed-source models with hidden biases making decisions that shape your worldview isn't a feature, it's a risk. Layer three: compute. Amazon may be picking winners in AI by controlling compute allocation. Cloud companies stopped being neutral marketplaces and started giving GPUs to the companies they've bet on. Amazon overwhelmingly backing Anthropic might have contributed to Claude's dominance. The cloud is no longer a utility. It's a kingmaker. Greg Osuri 🇺🇸's conclusion: intelligence should be free like water and air. Decentralized AI isn't ideology. It's the only architecture where the model can't ban your tools, the biases can be audited, and the compute isn't allocated by a company picking favorites.

The Rollup

15,005 görüntüleme • 4 ay önce

Rep. Jimmy Gomez — founder of the “Dads Caucus,” married with a baby at home — caught making out with Eric Swalwell’s chief of staff Yardena Wolf outside a party at Swalwell’s own place. Now Gomez is out here admitting his “personal mistakes” like it’s no big deal. Meanwhile Swalwell is drowning in his own sexual assault allegations and decades of sleazy behavior. These two California Democrats are the ultimate hypocrites, preaching family values and public service while treating Congress like their personal hookup playground. Birds of a feather, absolute clowns 🤡 who embody everything wrong with DC elitists — zero integrity, zero shame, and total betrayal of the people they claim to represent. The swamp never drains when these clowns keep flooding it with their drama. What should politically happen to Gomez? He should resign immediately or face a full House Ethics expulsion vote, get primaried hard by his own district, and be stripped of any committee assignments. Family-values frauds like this don’t deserve to keep their seats. Do you think this punishment would be appropriate? (Video: AI)
0:20

Rep. Jimmy Gomez — founder of the “Dads Caucus,” married with a baby at home — caught making out with Eric Swalwell’s chief of staff Yardena Wolf outside a party at Swalwell’s own place. Now Gomez is out here admitting his “personal mistakes” like it’s no big deal. Meanwhile Swalwell is drowning in his own sexual assault allegations and decades of sleazy behavior. These two California Democrats are the ultimate hypocrites, preaching family values and public service while treating Congress like their personal hookup playground. Birds of a feather, absolute clowns 🤡 who embody everything wrong with DC elitists — zero integrity, zero shame, and total betrayal of the people they claim to represent. The swamp never drains when these clowns keep flooding it with their drama. What should politically happen to Gomez? He should resign immediately or face a full House Ethics expulsion vote, get primaried hard by his own district, and be stripped of any committee assignments. Family-values frauds like this don’t deserve to keep their seats. Do you think this punishment would be appropriate? (Video: AI)

Paul A. Szypula 🇺🇸

43,057 görüntüleme • 2 ay önce

SORA + GLM-4.6 is next-level 🔥 This duo is insane for producing cinematic, high-quality videos no studios, no teams, no $1K/hour directors. Just two AI tools working together like a full creative department. Here’s how it works: → GLM-4.6 expands your concept into full narratives, detailed character actions, and 5–10 scene prompts → Sora converts GLM-4.6’s prompts into several video versions with perfect visuals, sound, and transitions → GLM-4.6 reviews each clip, checks consistency, improves pacing, and suggests refinements → Sora regenerates the enhanced scenes automatically Ideal for creators and brands testing cinematic content at scale without spending weeks or thousands. The magic is in the combo: GLM-4.6 = story depth + intelligent creative direction Sora = production quality that rivals professional studios While agencies take months, you’re generating 20–30 polished variations in a single afternoon. Test multiple angles. Refine fast. Scale smarter. Like, RT + reply with “SORA” and I’ll DM the full workflow (Must be following so I can DM). Skip this and keep paying agencies $15K/month for a handful of videos.
0:20

SORA + GLM-4.6 is next-level 🔥 This duo is insane for producing cinematic, high-quality videos no studios, no teams, no $1K/hour directors. Just two AI tools working together like a full creative department. Here’s how it works: → GLM-4.6 expands your concept into full narratives, detailed character actions, and 5–10 scene prompts → Sora converts GLM-4.6’s prompts into several video versions with perfect visuals, sound, and transitions → GLM-4.6 reviews each clip, checks consistency, improves pacing, and suggests refinements → Sora regenerates the enhanced scenes automatically Ideal for creators and brands testing cinematic content at scale without spending weeks or thousands. The magic is in the combo: GLM-4.6 = story depth + intelligent creative direction Sora = production quality that rivals professional studios While agencies take months, you’re generating 20–30 polished variations in a single afternoon. Test multiple angles. Refine fast. Scale smarter. Like, RT + reply with “SORA” and I’ll DM the full workflow (Must be following so I can DM). Skip this and keep paying agencies $15K/month for a handful of videos.

Md Riyazuddin

81,502 görüntüleme • 10 ay önce

$10,000 invested with Jim Cramer in 2001 turned into $28,997 by 2020. The same $10,000 in a free S&P 500 index fund? $44,957. Jim Cramer's CNBC Investing Club is TRASH... But they're still gassing him up. Wharton School researchers tracked Cramer's Action Alerts PLUS portfolio for over 17 years. The verdict: 3.38% annualized returns versus the S&P 500's 5.59%. More volatility. Lower returns. A Sharpe ratio of 0.11 versus the index's 0.24. You were taking on MORE risk for LESS reward. But it gets even worse. From inception through June 2023, cumulative total return: 230%. The S&P 500 with dividends reinvested: 453%. From 2016 to 2022, Action Alerts outperformed the market exactly ONCE. And you were paying $399 a year for the "privilege". Here's what really bothers me: CNBC sells this product to retail investors who trust the brand. Regular people. People without decades of experience to know better. They watch the flashing lights, the sound effects, the screaming - and assume there's actual substance behind this. There isn't. After adjusting for market risk, Cramer's excess return was "essentially zero." Not alpha. Noise. And SVB? First Republic? February 2023: Cramer told viewers to BUY Silicon Valley Bank. One month later it collapsed. He called First Republic "a very good bank." It failed two months after that. Catastrophic calls on systemically important institutions. Delivered to retail investors with total confidence. I've spent 45 years in this business. I worked for Peter Lynch. I know what genuine investment analysis looks like. What CNBC sells is infotainment dressed as financial advice. The index fund is free. It outperforms. Remember that the next time someone tries to sell you a $399 subscription to underperform the market. Say NO to Cramer and CNBC.
1:33

$10,000 invested with Jim Cramer in 2001 turned into $28,997 by 2020. The same $10,000 in a free S&P 500 index fund? $44,957. Jim Cramer's CNBC Investing Club is TRASH... But they're still gassing him up. Wharton School researchers tracked Cramer's Action Alerts PLUS portfolio for over 17 years. The verdict: 3.38% annualized returns versus the S&P 500's 5.59%. More volatility. Lower returns. A Sharpe ratio of 0.11 versus the index's 0.24. You were taking on MORE risk for LESS reward. But it gets even worse. From inception through June 2023, cumulative total return: 230%. The S&P 500 with dividends reinvested: 453%. From 2016 to 2022, Action Alerts outperformed the market exactly ONCE. And you were paying $399 a year for the "privilege". Here's what really bothers me: CNBC sells this product to retail investors who trust the brand. Regular people. People without decades of experience to know better. They watch the flashing lights, the sound effects, the screaming - and assume there's actual substance behind this. There isn't. After adjusting for market risk, Cramer's excess return was "essentially zero." Not alpha. Noise. And SVB? First Republic? February 2023: Cramer told viewers to BUY Silicon Valley Bank. One month later it collapsed. He called First Republic "a very good bank." It failed two months after that. Catastrophic calls on systemically important institutions. Delivered to retail investors with total confidence. I've spent 45 years in this business. I worked for Peter Lynch. I know what genuine investment analysis looks like. What CNBC sells is infotainment dressed as financial advice. The index fund is free. It outperforms. Remember that the next time someone tries to sell you a $399 subscription to underperform the market. Say NO to Cramer and CNBC.

George Noble

50,661 görüntüleme • 5 ay önce

Anyone with 8GB or 12GB VRAM setups needs to understand that "-ncmoe" is the key flag to boost performance on llama.cpp Here are my results for Qwen3.6 35B A3B, with 64k q8_0 context on a 8GB RTX 3070Ti: ⚪️ no flag → 8.7 tok/s RAM: 13.6GB & VRAM: 7.8GB 🔴 -ncmoe 35 → 27.5 tok/s RAM: 12.1GB & VRAM: 4.3GB 🟢 -ncmoe 30 → 32.5 tok/s RAM: 12GB & VRAM: 5.6GB 🔵 -ncmoe 25 → 40.9 tok/s RAM: 12GB & VRAM: 6.9GB Please note the ram and vram usage you see are total usage of a windows pc, with the model running. My friend's setup: 8GB VRAM and 16GB RAM. You can boost performance by switching to Linux, just something to keep in mind. Basically, this flag keeps the MoE experts in the first X layers on your CPU + RAM, instead of eating all your VRAM straight away. This is a smart hybrid offload way that lets you run bigger models without OOM while keeping the rest on your GPU for speed. As we can see on the data, there's a sweet spot. When we lower it from 35 to 25, speed bumps +50% because there are more layers on your GPU (look at the VRAM usage). The key here is to play around with the number and fit as much as possible on your VRAM, goal is to have 1GB/800MB headroom to avoid stress. ↓ server flags below
0:37

Anyone with 8GB or 12GB VRAM setups needs to understand that "-ncmoe" is the key flag to boost performance on llama.cpp Here are my results for Qwen3.6 35B A3B, with 64k q8_0 context on a 8GB RTX 3070Ti: ⚪️ no flag → 8.7 tok/s RAM: 13.6GB & VRAM: 7.8GB 🔴 -ncmoe 35 → 27.5 tok/s RAM: 12.1GB & VRAM: 4.3GB 🟢 -ncmoe 30 → 32.5 tok/s RAM: 12GB & VRAM: 5.6GB 🔵 -ncmoe 25 → 40.9 tok/s RAM: 12GB & VRAM: 6.9GB Please note the ram and vram usage you see are total usage of a windows pc, with the model running. My friend's setup: 8GB VRAM and 16GB RAM. You can boost performance by switching to Linux, just something to keep in mind. Basically, this flag keeps the MoE experts in the first X layers on your CPU + RAM, instead of eating all your VRAM straight away. This is a smart hybrid offload way that lets you run bigger models without OOM while keeping the rest on your GPU for speed. As we can see on the data, there's a sweet spot. When we lower it from 35 to 25, speed bumps +50% because there are more layers on your GPU (look at the VRAM usage). The key here is to play around with the number and fit as much as possible on your VRAM, goal is to have 1GB/800MB headroom to avoid stress. ↓ server flags below

left curve dev

166,368 görüntüleme • 3 ay önce

Deepseek V4 Flash 0731 (Q2) - 12 tokens/sec - Single RTX 4090 - 650+ tokens/sec prefill - 250k context - no kv cache quantization! DeepSeek just dropped the official V4 Flash 0731 two days ago with a massive agent capabilities upgrade. The official benchmarks are literally crushing their own V4-Pro-Preview on agentic tasks like Terminal Bench 2.1 and DeepSWE. Unsloth AI said they couldn't wait to bring it to local devices, and they delivered. If you thought my 118B Poolside Laguna S 2.1 MoE run last week on a single GPU was wild, hold onto your hardware. I just successfully ran Unsloth’s brand new 91GB DeepSeek-V4-Flash-0731 (UD-IQ2_M) GGUF entirely locally. And I pushed it to a mind-bending 250,000 context window. The VRAM ceiling is an illusion if you know how to optimize llama.cpp. Here are the benchmarks and the cheat codes to run a local frontier class model yourself. For the hardware and setup, I used a single NVIDIA RTX 4090 (24GB VRAM) hooked up via a PCIe 4 bus, running Ubuntu 22.04 LTS and CUDA 13.0. You don't need a massive enterprise server for this, if you have more than 80 GB of standard DDR4 RAM and a 24GB card like an RTX 3090 or 4090, you can run this exact stack yourself. All benchmarks were run using a massive 28k token prompt to truly stress test the prefill limits. no kv cache quantization THE BENCHMARKS (Scaling Context): # 80k Context (Baseline: -b 2048 -ub 2048): Prefill: 465.43 t/s | Decode: 13.00 t/s | VRAM: 22.87 GB # 80k Context (Optimized: -b 4096 -ub 4096): Prefill: 643.15 t/s | Decode: 12.20 t/s | VRAM: 23.00 GB (Notice how doubling the batch flags spiked my prefill throughput by nearly 200 t/s with almost zero VRAM penalty) # 180k Context (-b 4096 -ub 4096): Prefill: 629.18 t/s | Decode: 11.92 t/s | VRAM: 23.40 GB # 250k Context MAXIMUM (-b 4096 -ub 4096): Prefill: 619.02 t/s | Decode: 11.54 t/s | VRAM: 23.40 GB # THE SECRET SAUCE (Why this works): Unsloth’s UD-IQ2_M quant is ~91GB across 3 files. Since I only have 24GB of VRAM, the PCIe 4 bus and system RAM have to do the heavy lifting. The magic bullet is the --no-mmap flag. By completely bypassing OS disk paging, I forced llama.cpp to load the massive model weights directly into the system RAM upfront. Combined with Flash Attention (-fa on) and exactly 12 CPU threads (--threads 12), I maintained an incredibly stable 11.5+ tokens/sec decode speed even at a quarter million token context. # THE EXACT COMMAND: ./build/bin/llama-server -m /workspace/models/DeepSeek-V4-Flash-0731-UD-IQ2_M-00001-of-00003.gguf -c 250000 -fa on --port 8080 --threads 12 -b 4096 -ub 4096 --no-mmap -v Local conversational and agentic coding AI is fully here. You don’t need an API or an H100 cluster. Qwen 3.8 27b drops next week making the 24GB VRAM tier even more worthwhile. What does your current local AI rig look like, and what's the craziest model you've managed to squeeze into it? Official huggingface GGUF links from Unsloth and performance graphs are dropped in the replies below!
0:21

Deepseek V4 Flash 0731 (Q2) - 12 tokens/sec - Single RTX 4090 - 650+ tokens/sec prefill - 250k context - no kv cache quantization! DeepSeek just dropped the official V4 Flash 0731 two days ago with a massive agent capabilities upgrade. The official benchmarks are literally crushing their own V4-Pro-Preview on agentic tasks like Terminal Bench 2.1 and DeepSWE. Unsloth AI said they couldn't wait to bring it to local devices, and they delivered. If you thought my 118B Poolside Laguna S 2.1 MoE run last week on a single GPU was wild, hold onto your hardware. I just successfully ran Unsloth’s brand new 91GB DeepSeek-V4-Flash-0731 (UD-IQ2_M) GGUF entirely locally. And I pushed it to a mind-bending 250,000 context window. The VRAM ceiling is an illusion if you know how to optimize llama.cpp. Here are the benchmarks and the cheat codes to run a local frontier class model yourself. For the hardware and setup, I used a single NVIDIA RTX 4090 (24GB VRAM) hooked up via a PCIe 4 bus, running Ubuntu 22.04 LTS and CUDA 13.0. You don't need a massive enterprise server for this, if you have more than 80 GB of standard DDR4 RAM and a 24GB card like an RTX 3090 or 4090, you can run this exact stack yourself. All benchmarks were run using a massive 28k token prompt to truly stress test the prefill limits. no kv cache quantization THE BENCHMARKS (Scaling Context): # 80k Context (Baseline: -b 2048 -ub 2048): Prefill: 465.43 t/s | Decode: 13.00 t/s | VRAM: 22.87 GB # 80k Context (Optimized: -b 4096 -ub 4096): Prefill: 643.15 t/s | Decode: 12.20 t/s | VRAM: 23.00 GB (Notice how doubling the batch flags spiked my prefill throughput by nearly 200 t/s with almost zero VRAM penalty) # 180k Context (-b 4096 -ub 4096): Prefill: 629.18 t/s | Decode: 11.92 t/s | VRAM: 23.40 GB # 250k Context MAXIMUM (-b 4096 -ub 4096): Prefill: 619.02 t/s | Decode: 11.54 t/s | VRAM: 23.40 GB # THE SECRET SAUCE (Why this works): Unsloth’s UD-IQ2_M quant is ~91GB across 3 files. Since I only have 24GB of VRAM, the PCIe 4 bus and system RAM have to do the heavy lifting. The magic bullet is the --no-mmap flag. By completely bypassing OS disk paging, I forced llama.cpp to load the massive model weights directly into the system RAM upfront. Combined with Flash Attention (-fa on) and exactly 12 CPU threads (--threads 12), I maintained an incredibly stable 11.5+ tokens/sec decode speed even at a quarter million token context. # THE EXACT COMMAND: ./build/bin/llama-server -m /workspace/models/DeepSeek-V4-Flash-0731-UD-IQ2_M-00001-of-00003.gguf -c 250000 -fa on --port 8080 --threads 12 -b 4096 -ub 4096 --no-mmap -v Local conversational and agentic coding AI is fully here. You don’t need an API or an H100 cluster. Qwen 3.8 27b drops next week making the 24GB VRAM tier even more worthwhile. What does your current local AI rig look like, and what's the craziest model you've managed to squeeze into it? Official huggingface GGUF links from Unsloth and performance graphs are dropped in the replies below!

Alok

42,894 görüntüleme • 8 gün önce