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。

Gemma 4 E4B is impressive for an on-device LLM. GPT-4ish quality, and expect hallucinations. Here is: “List five sociological theories starting with u and what they are. Then describe them in a rhyming verse” Its in real time, the last is a little bit of a stretch, but not bad!

Ethan Mollick

364,443 subscribers

54,463 次观看 • 3 个月前 •via X (Twitter)

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

A long time coming but new mlx-lm is here with better batching support in the server and Gemma 4. pip install -U mlx-lm Here is a video where a single M3 Ultra serves 5 opencode sessions with Gemma 4 26B that process ~130k tokens in ~1.5 minutes.

A long time coming but new mlx-lm is here with better batching support in the server and Gemma 4. pip install -U mlx-lm Here is a video where a single M3 Ultra serves 5 opencode sessions with Gemma 4 26B that process ~130k tokens in ~1.5 minutes.

Angelos Katharopoulos

66,128 次观看 • 3 个月前

Watch Gemma-4-E4B casually identify sea animals by classifying images in a single agentic session using its vision capabilities. (impressive for a 4B model 🚀) I'm convinced: the agents of tomorrow are local, free, fast, and run on every computer!

Watch Gemma-4-E4B casually identify sea animals by classifying images in a single agentic session using its vision capabilities. (impressive for a 4B model 🚀) I'm convinced: the agents of tomorrow are local, free, fast, and run on every computer!

Victor M

171,280 次观看 • 3 个月前

to run gemma 4 on your phone without any login or internet: - download Google AI Edge Gallery - select Gemma 4 E2B/E4B - thats it I wish they give support to add other edge models in future like locallyAI here is a video of gemma 4 E4B in action at 13 tps on iphone 15 pro max

to run gemma 4 on your phone without any login or internet: - download Google AI Edge Gallery - select Gemma 4 E2B/E4B - thats it I wish they give support to add other edge models in future like locallyAI here is a video of gemma 4 E4B in action at 13 tps on iphone 15 pro max

Harveen Singh Chadha

13,552 次观看 • 3 个月前

Gameboy Player on Dolphin is so cool but this still is a bit undercooked in the current state, audio quality is kinda bad and changing roms is a bit annoying.

Gameboy Player on Dolphin is so cool but this still is a bit undercooked in the current state, audio quality is kinda bad and changing roms is a bit annoying.

Searcher

157,198 次观看 • 3 个月前

BARTIROMO: How much time is it gonna take? LUTNICK: We own the economy in the fourth quarter. First and second quarter are sort of a little bit our optimism ... but it is the mess that we were left with, but it's not that bad

BARTIROMO: How much time is it gonna take? LUTNICK: We own the economy in the fourth quarter. First and second quarter are sort of a little bit our optimism ... but it is the mess that we were left with, but it's not that bad

Aaron Rupar

690,649 次观看 • 1 年前

"This is how GPT-4 sees and hears itself" I used GPT-4 to describe itself. Then I used its description to generate an image, a video based on this image and a soundtrack. Tools I used: GPT-4, Midjourney, Kainber AI, Mubert, RunwayML This is the description I used that GPT-4 had of itself as a prompt to text-to-image, image-to-video, and text-to-music. I put the video and sound together in RunwayML.

"This is how GPT-4 sees and hears itself" I used GPT-4 to describe itself. Then I used its description to generate an image, a video based on this image and a soundtrack. Tools I used: GPT-4, Midjourney, Kainber AI, Mubert, RunwayML This is the description I used that GPT-4 had of itself as a prompt to text-to-image, image-to-video, and text-to-music. I put the video and sound together in RunwayML.

Kris Kashtanova

1,233,426 次观看 • 3 年前

OpenAI Greg Brockman: GPT-5.5 is not an endpoint — it's a beginning It is an early step toward stronger models arriving in the coming months, with larger gains across many capabilities The focus is not just on better benchmarks, but usefulness in the real world, for real users and real-world work

OpenAI Greg Brockman: GPT-5.5 is not an endpoint — it's a beginning It is an early step toward stronger models arriving in the coming months, with larger gains across many capabilities The focus is not just on better benchmarks, but usefulness in the real world, for real users and real-world work

Haider.

40,201 次观看 • 2 个月前

I've made a ton of money helping companies implement LLM-as-a-judge evaluations. LLM Judges provide a ton of value. But the hard part is choosing the model to implement the judge. • The family of GPT-5 models is very good, but slow and expensive. • Models like Gemma and Phi are fast and cheap, but not that good. Most of the time, you can only run a percentage of your traffic through the model (otherwise it would be too expensive and slow). But now, there's a better strategy.

I've made a ton of money helping companies implement LLM-as-a-judge evaluations. LLM Judges provide a ton of value. But the hard part is choosing the model to implement the judge. • The family of GPT-5 models is very good, but slow and expensive. • Models like Gemma and Phi are fast and cheap, but not that good. Most of the time, you can only run a percentage of your traffic through the model (otherwise it would be too expensive and slow). But now, there's a better strategy.

Santiago

29,916 次观看 • 2 个月前

🚨The Reptilians, Do they all have a Bad agenda for Humans?🚨 "Well, they look like you and I, but when you, look them in the eye and they get a little angry at you, they're, they get a little slit eye and that slit eye is like, oh, okay, there's a reptilian there. " "But not all reptilians are bad like we talk about, and not all, Draco Reptilian are bad."

🚨The Reptilians, Do they all have a Bad agenda for Humans?🚨 "Well, they look like you and I, but when you, look them in the eye and they get a little angry at you, they're, they get a little slit eye and that slit eye is like, oh, okay, there's a reptilian there. " "But not all reptilians are bad like we talk about, and not all, Draco Reptilian are bad."

UAPWixy

81,669 次观看 • 3 个月前

When a corporation is scared of its workers having a voice, you know the system is broken. It's time for Starbucks to step up. The baristas here in Gilbert are fighting for what they’ve earned and I’m proud to stand with them

When a corporation is scared of its workers having a voice, you know the system is broken. It's time for Starbucks to step up. The baristas here in Gilbert are fighting for what they’ve earned and I’m proud to stand with them

Ruben Gallego

16,150 次观看 • 7 个月前

Miami is not a real place for people thinking they are going to move there and live like they are on Vacation everyday. It’s a playground for the rich and most people don’t go to the beach everyday. As a matter of fact, I’d argue middle class in Miami is a lower quality of life than in a smaller city. What do you think?

Miami is not a real place for people thinking they are going to move there and live like they are on Vacation everyday. It’s a playground for the rich and most people don’t go to the beach everyday. As a matter of fact, I’d argue middle class in Miami is a lower quality of life than in a smaller city. What do you think?

Anton Daniels

92,847 次观看 • 1 年前

This is a trillion-dollar industry, and you can't solve it with an LLM: • Forecasting • Fraud detection • Churn prediction Large Language Models are fundamentally bad at solving these problems. When you feed structured data into an LLM, it doesn't see relationships, and it treats every number, date, and foreign key as a token. That's why you always get garbage back. An LLM thinks your database is a Wikipedia article. It doesn't understand its structure or its relationships. GPT-4 scores 63% on relational prediction tasks. That's the best it can do, and that's pretty much useless. You can't expect real-world business value to come from summarizing Wikipedia articles.

This is a trillion-dollar industry, and you can't solve it with an LLM: • Forecasting • Fraud detection • Churn prediction Large Language Models are fundamentally bad at solving these problems. When you feed structured data into an LLM, it doesn't see relationships, and it treats every number, date, and foreign key as a token. That's why you always get garbage back. An LLM thinks your database is a Wikipedia article. It doesn't understand its structure or its relationships. GPT-4 scores 63% on relational prediction tasks. That's the best it can do, and that's pretty much useless. You can't expect real-world business value to come from summarizing Wikipedia articles.

Santiago

94,701 次观看 • 2 个月前

President Trump is on The Five in just a few minutes and Greg is telling you what to expect

President Trump is on The Five in just a few minutes and Greg is telling you what to expect

The Five

46,131 次观看 • 3 个月前

Cursor CEO Michael Truell: Programming in the future might not mean programming with an AI helper, but in a new language entirely. "There is a future version of the world where the way you interact with AI is a little bit less like it's a human helper that you're delegating work to... and instead, it's a little bit more of an advance in compiler or interpreter technology." "It can lead you to a world where programming languages actually change. They can start to be a little bit more about what you want and a little bit less about how you do it." "There is a world where direct manipulation of the UI starts to play a little bit more into it." Michael Truell with Patrick Collison

Cursor CEO Michael Truell: Programming in the future might not mean programming with an AI helper, but in a new language entirely. "There is a future version of the world where the way you interact with AI is a little bit less like it's a human helper that you're delegating work to... and instead, it's a little bit more of an advance in compiler or interpreter technology." "It can lead you to a world where programming languages actually change. They can start to be a little bit more about what you want and a little bit less about how you do it." "There is a world where direct manipulation of the UI starts to play a little bit more into it." Michael Truell with Patrick Collison

a16z

92,499 次观看 • 4 个月前

Increasingly finding that the one thing that most makes managers panic about AI is showing them, not the advanced features of GPT-4, but rather the fact that Copilot for Office can create an OK PowerPoint with speaker notes from a document in 47 seconds. This is all real time.

Increasingly finding that the one thing that most makes managers panic about AI is showing them, not the advanced features of GPT-4, but rather the fact that Copilot for Office can create an OK PowerPoint with speaker notes from a document in 47 seconds. This is all real time.

Ethan Mollick

570,252 次观看 • 2 年前

At this point I’m starting to feel bad for beating a dead horse - but here is Sam Harris just a couple days ago claiming the left is fundamentally moral and the right has no standards at all in discussion with Preet Bharara. Delusion is fascinating to watch in real time.

At this point I’m starting to feel bad for beating a dead horse - but here is Sam Harris just a couple days ago claiming the left is fundamentally moral and the right has no standards at all in discussion with Preet Bharara. Delusion is fascinating to watch in real time.

FischerKing

135,104 次观看 • 1 年前

Talarico: "This is proof that there is something happening in Texas tonight tonight, the people of our state gave this country a little bit of hope, and a little bit of hope is a dangerous thing." He then walks off to Bad Bunny's "DTMF"

Talarico: "This is proof that there is something happening in Texas tonight tonight, the people of our state gave this country a little bit of hope, and a little bit of hope is a dangerous thing." He then walks off to Bad Bunny's "DTMF"

nikki mccann ramírez

734,751 次观看 • 4 个月前

Our biggest worry is mum sleeping on one and squashing them. Thats a real possibility when they are so small. Now that they are a little bigger they can sort of wiggle around and free themselves up. Someone has been in watching her and helping most of the time (4/7)

Our biggest worry is mum sleeping on one and squashing them. Thats a real possibility when they are so small. Now that they are a little bigger they can sort of wiggle around and free themselves up. Someone has been in watching her and helping most of the time (4/7)

Niall Harbison

105,354 次观看 • 1 年前

"What is going on with the younger generation of men?" "They are not men!" "Saw a video of a guy who said he wakes up at 5AM and makes an iced coffee. Dude the sun is up!" "Then he has a skin routine. Who has a skin routine? No you don't you swamp donkey!" "I wake up early and the first thing I do is scratch my nuts." That last bit had my cracking up! He ain't wrong! This younger generation of men are hyper feminized, not all, but a lot of them!

"What is going on with the younger generation of men?" "They are not men!" "Saw a video of a guy who said he wakes up at 5AM and makes an iced coffee. Dude the sun is up!" "Then he has a skin routine. Who has a skin routine? No you don't you swamp donkey!" "I wake up early and the first thing I do is scratch my nuts." That last bit had my cracking up! He ain't wrong! This younger generation of men are hyper feminized, not all, but a lot of them!

AmericanPapaBear™

273,107 次观看 • 7 个月前

streets are saying they were in a hotel last time, and this is what they are doing:

streets are saying they were in a hotel last time, and this is what they are doing:

jiji

28,086 次观看 • 6 个月前