正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

This is very likely an Anthropic model. It has the exact same context 200k window as Sonnet, its SVG output styles are literally the same as Claude. Its tokens per second are not fast enough to be Haiku. (Could be intentionally slow)

JB

4,685 subscribers

173,951 次观看 • 6 个月前 •via X (Twitter)

新闻政治教育科学技术

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

Claude Sonnet 4.8 is imminent. Most likely dropping today. Sonnet 4.6 is starting to look outdated. Look at the Flappy Bird test on BridgeBench. GPT 5.5 and Claude Opus 4.7 have moved past it. The model that used to be the default for most vibe coders is falling behind. Anthropic needs Sonnet 4.8 to close the gap. The second it drops, we're running it through every BridgeBench test and posting the results live. New Sonnet day. Let's see what Anthropic has.

Claude Sonnet 4.8 is imminent. Most likely dropping today. Sonnet 4.6 is starting to look outdated. Look at the Flappy Bird test on BridgeBench. GPT 5.5 and Claude Opus 4.7 have moved past it. The model that used to be the default for most vibe coders is falling behind. Anthropic needs Sonnet 4.8 to close the gap. The second it drops, we're running it through every BridgeBench test and posting the results live. New Sonnet day. Let's see what Anthropic has.

Bridgebench

16,715 次观看 • 2 个月前

Your Claude has a tell, and almost nobody uses it. Make it start every reply with your name. One line in custom instructions, no code. The moment it stops saying your name, its context window is full. That is the exact point the output starts drifting, and you do not notice. Name stops? Start a fresh chat. It has killed about 90% of my “why has Claude gone weird” moments. Miners sent a canary down the pit to catch the gas they could not see. Same trick, for your context. Comment GUIDE and I’ll send you the exact canary setup.

Your Claude has a tell, and almost nobody uses it. Make it start every reply with your name. One line in custom instructions, no code. The moment it stops saying your name, its context window is full. That is the exact point the output starts drifting, and you do not notice. Name stops? Start a fresh chat. It has killed about 90% of my “why has Claude gone weird” moments. Miners sent a canary down the pit to catch the gas they could not see. Same trick, for your context. Comment GUIDE and I’ll send you the exact canary setup.

Charlie Hills

24,095 次观看 • 29 天前

Ben Davis thinks the Phillies are likely to run it back, but is worried it will be the same story: “Their offense is just not good enough.”

Ben Davis thinks the Phillies are likely to run it back, but is worried it will be the same story: “Their offense is just not good enough.”

SPORTSRADIO 94WIP

21,798 次观看 • 9 个月前

Fast tokens are the most valuable tokens. Because inference speed is productivity. And Cerebras makes the fastest tokens in the world. Google DeepMind's Gemma 4 runs at 1,500+ tokens per second on Cerebras — more than an order of magnitude faster than on GPUs and 15x faster than Claude Haiku 4.5. Private preview today. GA later this month. Build blazing fast multimodal apps.

Fast tokens are the most valuable tokens. Because inference speed is productivity. And Cerebras makes the fastest tokens in the world. Google DeepMind's Gemma 4 runs at 1,500+ tokens per second on Cerebras — more than an order of magnitude faster than on GPUs and 15x faster than Claude Haiku 4.5. Private preview today. GA later this month. Build blazing fast multimodal apps.

Andrew Feldman

12,008 次观看 • 1 个月前

We are not the same! This is not working! We should not be forced to live this way! It’s all going to end very badly as the British people are suffering and we don’t deserve this!

We are not the same! This is not working! We should not be forced to live this way! It’s all going to end very badly as the British people are suffering and we don’t deserve this!

Alexandra

36,807 次观看 • 3 个月前

🚨 BITCOIN IS APPROACHING THE BULL RUN LEVEL THAT NEVER FAILED Throughout history, the same exact trendline has marked the bottom for Bitcoin. We are inches away from that exact same trendline as we speak. According to this model, the top has not been put in yet. Once Bitcoin retests that trendline, the top should be put in at around $500,000. Prepare for the largest bull run to date. Most people are going to miss this. Don't be one of those people.

🚨 BITCOIN IS APPROACHING THE BULL RUN LEVEL THAT NEVER FAILED Throughout history, the same exact trendline has marked the bottom for Bitcoin. We are inches away from that exact same trendline as we speak. According to this model, the top has not been put in yet. Once Bitcoin retests that trendline, the top should be put in at around $500,000. Prepare for the largest bull run to date. Most people are going to miss this. Don't be one of those people.

Coinvo Trading

23,636 次观看 • 5 个月前

Got an output from the strongest pre-release Gemini 3.5 Pro checkpoint 👀 I compared it to Claude Fable 5 and GPT-5.5 on the same exact prompt Yeah... This model is mostly dead on arrival (tried it on other prompts too ofc). It will likely be the "poor man's model" next to 3.5 flash. 1st video: Gemini 3.5 Pro (pre-release) 2nd video: GPT-5.5 3rd video: Claude Fable 5

Got an output from the strongest pre-release Gemini 3.5 Pro checkpoint 👀 I compared it to Claude Fable 5 and GPT-5.5 on the same exact prompt Yeah... This model is mostly dead on arrival (tried it on other prompts too ofc). It will likely be the "poor man's model" next to 3.5 flash. 1st video: Gemini 3.5 Pro (pre-release) 2nd video: GPT-5.5 3rd video: Claude Fable 5

Lentils

50,395 次观看 • 2 个月前

China's Qafind Labs just launched the first Diffusion Language Model (DLM). 🔥 ChatDLM, described as the first Diffusion Language Model (DLM), will be open-sourced soon. - Inference Speed: 2,800 tokens/sec (on A100), which is insanely fast. - Context Window: 131,072 tokens. More details 👇

China's Qafind Labs just launched the first Diffusion Language Model (DLM). 🔥 ChatDLM, described as the first Diffusion Language Model (DLM), will be open-sourced soon. - Inference Speed: 2,800 tokens/sec (on A100), which is insanely fast. - Context Window: 131,072 tokens. More details 👇

AshutoshShrivastava

73,296 次观看 • 1 年前

ANTHROPIC FOUND THE MOST EXPENSIVE PART OF AI, EVERY EMPLOYEE STARTS FROM ZERO. one person explains the project to Claude. the answer disappears inside a private chat. the next employee opens another Claude and repeats the same context. Claude Tag removes that reset. everyone in a Slack channel works with the same agent. It sees the thread, remembers relevant decisions, uses the tools assigned to that channel and leaves the output where the next person can continue. Anthropic says the internal version already creates 65% of its product team’s code. the model did not suddenly become 65% smarter. the context stopped disappearing between people.

ANTHROPIC FOUND THE MOST EXPENSIVE PART OF AI, EVERY EMPLOYEE STARTS FROM ZERO. one person explains the project to Claude. the answer disappears inside a private chat. the next employee opens another Claude and repeats the same context. Claude Tag removes that reset. everyone in a Slack channel works with the same agent. It sees the thread, remembers relevant decisions, uses the tools assigned to that channel and leaves the output where the next person can continue. Anthropic says the internal version already creates 65% of its product team’s code. the model did not suddenly become 65% smarter. the context stopped disappearing between people.

Myttle

30,351 次观看 • 1 个月前

man anthropic is absolutely cooking microsoft. claude now works natively in word, powerpoint and excel. thats 450M users but how the fuck did anthropic beat microsoft co-pilot? it gets worse - microsoft blew $13B creating co-pilot and owns 27% of openai yet no sign on chatgpt. - claude has access to SHARED context across excel, word and powerpoint. they know the footprint of microsoft's users - very valuable AI data. - only 15M microsoft users pay for copilot, so its clear they need to go model-agnostic - this is the 3rd major microsoft product to leverage claude (cowork and deep research are the other 2) very impressive execution from anthropic tbh, they are trying to be everywhere and anywhere - its working

man anthropic is absolutely cooking microsoft. claude now works natively in word, powerpoint and excel. thats 450M users but how the fuck did anthropic beat microsoft co-pilot? it gets worse - microsoft blew $13B creating co-pilot and owns 27% of openai yet no sign on chatgpt. - claude has access to SHARED context across excel, word and powerpoint. they know the footprint of microsoft's users - very valuable AI data. - only 15M microsoft users pay for copilot, so its clear they need to go model-agnostic - this is the 3rd major microsoft product to leverage claude (cowork and deep research are the other 2) very impressive execution from anthropic tbh, they are trying to be everywhere and anywhere - its working

Ejaaz

473,708 次观看 • 4 个月前

I remember watching this video for the first time thinking what it would be like when klopp leaves and its just the exact same as this so out the blue 👎🏻😢

I remember watching this video for the first time thinking what it would be like when klopp leaves and its just the exact same as this so out the blue 👎🏻😢

max penman

419,454 次观看 • 2 年前

We've been working hard to make Claude in Chrome as fast as possible. We are rolling out an experimental "Quick Mode" to all users today. It is particularly impressive when combined a fast model like Opus 4.6 (fast mode) or Haiku 4.5. I look forward to hearing what you think!

We've been working hard to make Claude in Chrome as fast as possible. We are rolling out an experimental "Quick Mode" to all users today. It is particularly impressive when combined a fast model like Opus 4.6 (fast mode) or Haiku 4.5. I look forward to hearing what you think!

Gabe Mulley

28,547 次观看 • 5 个月前

your agent reviewing its own work is not a check. it is a second opinion from the same source. this is the most common gap in agent systems and it hides in plain sight, because the step exists. there is a review. it just cannot do the thing you think it does. here is the mechanism. the model produced an output from a context. you then ask the same model, holding the same context, whether that output is correct. it answers fluently, because that is what it does. and the answer is drawn from the same distribution that produced the thing being judged. same weights, same window, same blind spots. if the reason the output is wrong is something the model does not know, the review does not know it either. if the reason is something the context does not contain, the review has the same context. the failure mode and the detector share a cause. > why it feels like it works because most of the time the output is fine, and the review says fine. agreement is not evidence of detection. a reviewer that says pass on everything agrees with reality most of the time too. what you actually want to measure is what happens on the cases that are wrong. that is the only place a check earns its name, and it is exactly the place where a self-review is weakest. there is research on this. Huang and colleagues at DeepMind showed at ICLR 2024 that intrinsic self-correction, revising without external grounding, does not reliably help and often makes things worse. > what to actually do move the check outside the model. a test that runs, a schema that validates, a file that exists or does not, an exit code from something you did not write. these are not smarter than the model. they are just not correlated with it, and that is the entire value. when the judgement genuinely needs a model, at minimum use a different family. same family means shared blind spots, and frontier judges measurably inflate scores for outputs that look like their own. and split the work by kind. anything objectively checkable goes to code. only the genuinely semantic calls go to a judge, and those get a rubric written as one line. a review inside the loop tells you the model is confident. a check outside it tells you whether the work is done. save this - then read the eval setup below

your agent reviewing its own work is not a check. it is a second opinion from the same source. this is the most common gap in agent systems and it hides in plain sight, because the step exists. there is a review. it just cannot do the thing you think it does. here is the mechanism. the model produced an output from a context. you then ask the same model, holding the same context, whether that output is correct. it answers fluently, because that is what it does. and the answer is drawn from the same distribution that produced the thing being judged. same weights, same window, same blind spots. if the reason the output is wrong is something the model does not know, the review does not know it either. if the reason is something the context does not contain, the review has the same context. the failure mode and the detector share a cause. > why it feels like it works because most of the time the output is fine, and the review says fine. agreement is not evidence of detection. a reviewer that says pass on everything agrees with reality most of the time too. what you actually want to measure is what happens on the cases that are wrong. that is the only place a check earns its name, and it is exactly the place where a self-review is weakest. there is research on this. Huang and colleagues at DeepMind showed at ICLR 2024 that intrinsic self-correction, revising without external grounding, does not reliably help and often makes things worse. > what to actually do move the check outside the model. a test that runs, a schema that validates, a file that exists or does not, an exit code from something you did not write. these are not smarter than the model. they are just not correlated with it, and that is the entire value. when the judgement genuinely needs a model, at minimum use a different family. same family means shared blind spots, and frontier judges measurably inflate scores for outputs that look like their own. and split the work by kind. anything objectively checkable goes to code. only the genuinely semantic calls go to a judge, and those get a rubric written as one line. a review inside the loop tells you the model is confident. a check outside it tells you whether the work is done. save this - then read the eval setup below

Hanako

13,683 次观看 • 5 天前

XDefiant has missed its ideal launch window and chance to capitalize on a seemingly very down year for COD player satisfaction. Do you think the game can still be successful if it releases close to the same time as the new Call of Duty? #MW3

XDefiant has missed its ideal launch window and chance to capitalize on a seemingly very down year for COD player satisfaction. Do you think the game can still be successful if it releases close to the same time as the new Call of Duty? #MW3

ModernWarzone

2,994,996 次观看 • 2 年前

#WATCH | Mumbai, Maharashtra | Cancer survivor and actor Hina Khan says, "Our PM has seen a dream and the way we are moving forward, it seems very soon we will be 'Viksit Bharat'." On the treatment of cancer, she says, "...The protocols on a global level are the same. The treatment is the same, and the medicines which are given are the same. I will be given the same medicines as a person who might not be that financially strong. There is no difference in medicines."

#WATCH | Mumbai, Maharashtra | Cancer survivor and actor Hina Khan says, "Our PM has seen a dream and the way we are moving forward, it seems very soon we will be 'Viksit Bharat'." On the treatment of cancer, she says, "...The protocols on a global level are the same. The treatment is the same, and the medicines which are given are the same. I will be given the same medicines as a person who might not be that financially strong. There is no difference in medicines."

ANI

92,306 次观看 • 1 年前

It has to be said, England fans are nowhere near as loud as the Scotland fans were during the anthems, in the same stadium… 🫣😬

It has to be said, England fans are nowhere near as loud as the Scotland fans were during the anthems, in the same stadium… 🫣😬

Football Away Days

760,182 次观看 • 1 个月前

Anthropic shipped 125 settings for Claude The official docs cover 40 One developer found the other 85 and his API bill dropped from $340 to $87 - not by using a cheaper model - not by writing shorter prompts just by moving one line in a config file to the right place > memory scoped per project → past clients never bleed into new work > Extended Thinking on Light by default → 18–25% fewer Opus tokens in week one > cache_control moved to the right line → the fix that turned a $340 bill into $87 > plugins and MCP servers toggled off when idle → saved 25–40K tokens per session > per-project model override → Haiku for docs, Sonnet for infra, Opus only where it matters most Claude users are running a $100/month tool at 30% of its actual capability

Anthropic shipped 125 settings for Claude The official docs cover 40 One developer found the other 85 and his API bill dropped from $340 to $87 - not by using a cheaper model - not by writing shorter prompts just by moving one line in a config file to the right place > memory scoped per project → past clients never bleed into new work > Extended Thinking on Light by default → 18–25% fewer Opus tokens in week one > cache_control moved to the right line → the fix that turned a $340 bill into $87 > plugins and MCP servers toggled off when idle → saved 25–40K tokens per session > per-project model override → Haiku for docs, Sonnet for infra, Opus only where it matters most Claude users are running a $100/month tool at 30% of its actual capability

Shadow Nick

379,914 次观看 • 2 个月前

like these voicelines are very unmistakably meant to be M2 considering the very apparent glitch effect that only the mimic has as well as it very obviously repeating things its heard for the most part so like. what do we make of this

like these voicelines are very unmistakably meant to be M2 considering the very apparent glitch effect that only the mimic has as well as it very obviously repeating things its heard for the most part so like. what do we make of this

EVA 🍉

41,167 次观看 • 1 年前