0:00 Intro 0:34 New OpenAI Model and Plan 2:58 OpenAI Fine-Tuning 4:39 Additional OpenAI Announcements 6:00 All The Google Announcements 9:52 Genie 2 - Generative Worlds 12:18 World Labs Generative Worlds 13:31 AI Store Builder 14:58 Grok For Free 15:49 Meta's New Model 16:19 Microsoft's Announcements 17:32 Amazon's Announcements 19:29 Luma's New Video Model 20:20 Hailuo's New Video Model 21:04 Tencent New Video Model 21:53 Leonardo Flow State 25:14 ElevenLabs Announcements 26:44 Hume Voice Control 27:48 Perplexity's New Partners 28:14 Elon Vs. OpenAI 30:10 New US "AI and Crypto Czar" 30:53 Humane's New OS 31:50 Clone Android 32:22 Final Thoughts
I am almost certain that every time an LLM releases a new version that reaches the top positions of any Benchmark, it is because it included in its dataset the questions and prompts specific to the benchmark.
@@vaibhavsingh630 chat-gpt is so extremely far ahead of the others...in % the others fell like 1-20% of the 100% chat-gpt has right now...its really ridicolous....no clue who makes their own % lists..
matt, i LOVE your ai news. always been into tech. recently (last month +) getting into ai for content creation and found myself down the rabbit hole. thank you for providing such value and entertainment!
Matt , you have been a staple in my weekly AI news. The Spotify one was amazing , in a huge notebook LLM fan. I continue to share your channel professionally and personally. Keep it up!
Hey Matt, wonderful channel, amazing content, and a great host, I love your excitement!!. I have been following for over a year (I think...maybe 2 :) now. On the Ai News videos it would be helpful to put chapters, I am not sure how difficult it is to set up but..., just a tip. Cheers, and continue the awesome work... as long as you still love it.
Excellent AI News this week Matt. Thank you so much. Can you do a FULL Grok demo please? Maybe a livestream with it? For Christmas. lol. That would be so much fun!!🤩
I like your content, sheer enthusiasm and how you break down issues. I am wondering whether you could once or twice a year expand your content beyond announcements, graphics generation, videos and voice to include how AI is being used in other domains. Possibly medical, maths, engineering etc. I realise that it may not ‘sell’ as well, but it would give your subscribers a broader perspective. I think this is an aspect of AI that is missing from TH-cam, but I stand to be corrected. Not a criticism but rather a request.
I am subbing to the pro mode so I can relentlessly prompt it, till Chatgpt designs a Time machine so I can go back and Invest all in Nvidia. Merry Xmas Matt!
13:20 How does Ai create full content with just one photo or image? I'm blown away 😲. 23:38 I like this expansive and scrolling approach to zero-in on what you want. 25:52 Having the option to use your own voice with the podcast generator tool is truly awesome. But I'm wondering if it is a suitable way to replace your real self to broadcast online. People might say that is cheating, cheating into doing less work by giving the public a replica of your voice (not the real you). Nevertheless I must acknowledge that it can be a great tool. 12:01 They are making it look so easy to simulate a virtual world with those Ai tools. I must say that the quality is highly good 👍🏾.
Thanks Matt. This week also brought an update to Suno v4.0, which made AI-generated music even more impressive, with significant improvements in vocals and song structure.
@0:00 Intro @0:34 New OpenAI Model and Plan @2:58 OpenAI Fine-Tuning @4:39 Additional OpenAI Announcements @6:00 All The Google Announcements @9:52 Genie 2 - Generative Worlds @12:18 World Labs Generative Worlds @13:31 AI Store Builder @14:58 Grok For Free @15:49 Meta's New Model @16:19 Microsoft's Announcements @17:32 Amazon's Announcements @19:29 Luma's New Video Model @20:20 Hailuo's New Video Model @21:04 Tencent New Video Model @21:53 Leonardo Flow State @25:14 ElevenLabs Announcements @26:44 Hume Voice Control @27:48 Perplexity's New Partners @28:14 Elon Vs. OpenAI @30:10 New US "AI and Crypto Czar" @30:53 Humane's New OS @31:50 Clone Android @32:22 Final Thoughts
Looking forward to when I can self-host these world building models and import them to engines like unity or whatever version Unreal Engine is on now; I imagine such a thing would be a boon for people who aren't artists but still want to build a small scale scene to explore.
@Matt, I asked Perplexity AI "Is there an AI model that can remove music from audio and video, leaving speech and other sounds?". Answer, "Yes, there are several AI models and tools that can remove music from audio and video while preserving speech and other sounds:", and a list where some are relevant. I haven't checked, so maybe you've tried and there's some reason for not using them.
The ideal marriage would be AI and quantum computing in order to create models that accurately predict not just the weather but also earth quakes. To me, that should be the priority everyone outs their efforts into.
the o1 is pretty limited in answers and when i want to write storys with it its super restricted so only up to 12 year apropriate stories....meanwhile when i see whats on the tv/youtube/movies its really facepalm...
Hey Matt! While you're out there wrapping up your Spotify year, there’s one huge update you totally missed. The whole NotebookLM crew just left Google this week to kick off something new. Get ready, this could be the next big thing in AI tools! 🤖👀🚀 Stay tuned!
My guess is immortality by 2045. Longevity escape velocity has already been happening for those who will live to see the day immortality becomes available and its definitely less than 75 years away.
Great Ai news. Open Ai Sora just seems to be all talk, and demos, while other video models have been running already. Amazon has also been sleeping and very slow to innovate via AWS. They just updated their terrible AWS console UIX. No major improvements compared to Google and Azure.
No. Both Pandora's box and the proverbial floodgates are open. The cat is out of the bag and the toothpaste is out of the tube. There's no going back. The future is here, the future is now.
You can physically see that this video was completely generated with AI by how many cuts and edits are made during video playback. Seems like the dead internet theory became reality.
the only update i want to hear is that these chatbots actually understand the words they are saying. until then like noam chomsky said 2 years ago about chatgpt and still remains true through now, is that these bot are just a parlor trick and bring us no closer to AI. As a computer scientist i could not agree more.
If we have human-sounding robots with artificial human-like organs then in 10 years there will for sure be real discussions of... uhmm... does a robot have to give consent.
Tron telemetric .... in the androidic consiones energetic constallations overlayerings levels of interaction, genuine computer accellerating towards the point were VR AR are being produced by genuine computer/s .... and not so much humans obtaining rightiones points for interaction in evolution creation...
Thanks Matt, nobody gives such clear relevant AI news like you ! Once again thank you!❤
Your cheerfulness is contagious.
Can you share what “supplements” you use? 😉
Thanks Matt. Your videos are always #1 on my watch list on AI. Peter
Thanks Matt, early subscriber here. Glad to see the evolution you're making. Congrats 🎉❤
0:00 Intro
0:34 New OpenAI Model and Plan
2:58 OpenAI Fine-Tuning
4:39 Additional OpenAI Announcements
6:00 All The Google Announcements
9:52 Genie 2 - Generative Worlds
12:18 World Labs Generative Worlds
13:31 AI Store Builder
14:58 Grok For Free
15:49 Meta's New Model
16:19 Microsoft's Announcements
17:32 Amazon's Announcements
19:29 Luma's New Video Model
20:20 Hailuo's New Video Model
21:04 Tencent New Video Model
21:53 Leonardo Flow State
25:14 ElevenLabs Announcements
26:44 Hume Voice Control
27:48 Perplexity's New Partners
28:14 Elon Vs. OpenAI
30:10 New US "AI and Crypto Czar"
30:53 Humane's New OS
31:50 Clone Android
32:22 Final Thoughts
Dude this was already there, what do you mean? I'm shocked that you even got 2 likes too
@@Someone7R7 your speaking into the void its a bot
@@gtaocinematics oh, it is just a stupid bot then, okay
I am almost certain that every time an LLM releases a new version that reaches the top positions of any Benchmark, it is because it included in its dataset the questions and prompts specific to the benchmark.
that what im saying. so from your point of you which is best LLm right now
@@vaibhavsingh630 The best one is the one you actually use
lmaoooo yeah...
@@vaibhavsingh630 chat-gpt is so extremely far ahead of the others...in % the others fell like 1-20% of the 100% chat-gpt has right now...its really ridicolous....no clue who makes their own % lists..
can't that equally decrease the quality?
matt, i LOVE your ai news. always been into tech. recently (last month +) getting into ai for content creation and found myself down the rabbit hole. thank you for providing such value and entertainment!
Matt , you have been a staple in my weekly AI news. The Spotify one was amazing , in a huge notebook LLM fan. I continue to share your channel professionally and personally. Keep it up!
Great to see you back Matt.
You say it almost every time, Matt! And you are absolutely right🤗
Great job, lots of information overload!
Matt is THE man!
You're a great resource.
What a comprehensive overview of new AI ! Thanks.
My mind has officially been blown! Not just because of what AI can do but the speed at which AI is moving!
Congratulations for all those videos. Always interesting!
Hey Matt, wonderful channel, amazing content, and a great host, I love your excitement!!. I have been following for over a year (I think...maybe 2 :) now. On the Ai News videos it would be helpful to put chapters, I am not sure how difficult it is to set up but..., just a tip. Cheers, and continue the awesome work... as long as you still love it.
Ups!, I just read the previous comment, sorry for saying... pretty much the same. Cheers
Excellent AI News this week Matt. Thank you so much. Can you do a FULL Grok demo please? Maybe a livestream with it? For Christmas. lol. That would be so much fun!!🤩
I like your content, sheer enthusiasm and how you break down issues. I am wondering whether you could once or twice a year expand your content beyond announcements, graphics generation, videos and voice to include how AI is being used in other domains. Possibly medical, maths, engineering etc. I realise that it may not ‘sell’ as well, but it would give your subscribers a broader perspective. I think this is an aspect of AI that is missing from TH-cam, but I stand to be corrected. Not a criticism but rather a request.
thx matt w.! 🙏🇺🇸
I am subbing to the pro mode so I can relentlessly prompt it, till Chatgpt designs a Time machine so I can go back and Invest all in Nvidia. Merry Xmas Matt!
merry Christmassss, AI just😂 took your job
Insane 🎉
13:20 How does Ai create full content with just one photo or image? I'm blown away 😲. 23:38 I like this expansive and scrolling approach to zero-in on what you want. 25:52 Having the option to use your own voice with the podcast generator tool is truly awesome. But I'm wondering if it is a suitable way to replace your real self to broadcast online. People might say that is cheating, cheating into doing less work by giving the public a replica of your voice (not the real you). Nevertheless I must acknowledge that it can be a great tool. 12:01 They are making it look so easy to simulate a virtual world with those Ai tools. I must say that the quality is highly good 👍🏾.
FINALLY a 60FPS thank you! Easy on the eyes!
So much going on. Thank you very much for keeping us up to date and for being all around awesome
Love your videos! Learning from them each time .
Question about gpt models which model is best practice for coding tasks?
Thanks Matt. This week also brought an update to Suno v4.0, which made AI-generated music even more impressive, with significant improvements in vocals and song structure.
Hopefully, Open AI will release the live video mode within the advance voice mode. Been waiting forever
A new map every single game is crazy
"Wildest week in AI news" heard that about 49 times this year lol
oh that leonardo update looks nice!!!
What are the costs involved for generating so many images. Does a monthly plan cover unlimited?
its mid im looking at it now its like bait to get u to buy a lot of features are off too
Seems on the free tier Grok adds a watermark in the lower right corner.😐
Elon's tramp stamp
Elon payed 44 Billion for Twitter, he has got to pay for that some how!
Grok sucks Elon fluffers...
@@fl1ntlock1 Grok isn't very good today. But I am looking forward to testing out Grok 3, which should be a good upgrade.
pretty sure you can find an ai service that removes watermark
FYI: To run Hunyuan Video models from Tencent, you need 45G to 60G of VRam.
80GB of memory for better generation quality.
Nonsense. You can already run Hunyuan Video models on literally 8GB VRAM.
@@TheGalacticIndian Great. I must of missed it. That is their requirements on Github page I went to. Have you tried it?
OpenAI and Sam Altman going all in with the Pentagon to develop weapons should have been the focus of this week's episode.
As an advisor to Leonardo, can you tell them to set the default to the new features page. Having to keep sliding the button on and off is annoying.
31:51 sox toy 🌚
Neesh as in- sheesh ya'll ! it's only an extra 0 .. I'm ordering 2 now !
So much stuff, so so much, it's impossible to try it all
What does Sam altman mean by think even harder? That bothers me so much because he didn't even explain the actual difference between them
Seems pretty obvious to me. Thinking harder just means more iterative refinement
There is more compute power for complex tasks.
@@carnivorous_vegan and that's what we call the Dunning Krueger effect, it could be majority voting, It could be many things, you have no idea lol
Have you tried asking Chat GTP?
@@countofst.germain6417 Those things you listed are iterative refinements.
Ok We definitely need AI-supported computers. So we can better things done. Can you create a video about it?
EVERY AI NEWS GOES LIKE - THIS HAS BEEN THE WILDEST WEEK SO FAR!!
Europe be like - We haven’t had a war since the last war!
Hey @Matt where would you say is a great resource to learn how to use AI for business, for solopreneurs?
Grok isn’t available for me, still have to subscribe. Is it that I am using an iPhone (not web) or that I live in Europe?
Thank you for correcting your pronunciation ❤
A VIDEO THIS LONG NEEDS TIMESTAMPS!
Why are shouting?
I totally agree
look at description lol
@0:00 Intro
@0:34 New OpenAI Model and Plan
@2:58 OpenAI Fine-Tuning
@4:39 Additional OpenAI Announcements
@6:00 All The Google Announcements
@9:52 Genie 2 - Generative Worlds
@12:18 World Labs Generative Worlds
@13:31 AI Store Builder
@14:58 Grok For Free
@15:49 Meta's New Model
@16:19 Microsoft's Announcements
@17:32 Amazon's Announcements
@19:29 Luma's New Video Model
@20:20 Hailuo's New Video Model
@21:04 Tencent New Video Model
@21:53 Leonardo Flow State
@25:14 ElevenLabs Announcements
@26:44 Hume Voice Control
@27:48 Perplexity's New Partners
@28:14 Elon Vs. OpenAI
@30:10 New US "AI and Crypto Czar"
@30:53 Humane's New OS
@31:50 Clone Android
@32:22 Final Thoughts
Newbie !? Please see the description so that you can get what you want !
Looking forward to when I can self-host these world building models and import them to engines like unity or whatever version Unreal Engine is on now; I imagine such a thing would be a boon for people who aren't artists but still want to build a small scale scene to explore.
Grok still paid version in UK
@Matt, I asked Perplexity AI "Is there an AI model that can remove music from audio and video, leaving speech and other sounds?". Answer, "Yes, there are several AI models and tools that can remove music from audio and video while preserving speech and other sounds:", and a list where some are relevant. I haven't checked, so maybe you've tried and there's some reason for not using them.
The ideal marriage would be AI and quantum computing in order to create models that accurately predict not just the weather but also earth quakes. To me, that should be the priority everyone outs their efforts into.
Is OpenAI's o1 model better at coding related task than Anthropic's Claude 3.5 Sonnet?
Why does the Leonardo AI image outputs still look like cartoonish clip art three years later? - RIP Leonardo AI
I'm enjoying the less clickbait titles😁
I'm in the UK - we don't have free GROK yet.... 🤔
The last time you said niche was the correct way.
the o1 is pretty limited in answers and when i want to write storys with it its super restricted so only up to 12 year apropriate stories....meanwhile when i see whats on the tv/youtube/movies its really facepalm...
How does someone submit AI news to you Matt?
Was that virtual Matt in spots❓❓
My ❤
Sooo…. “Flow State” seems like an almost identical concept to Decohere’s stations which have been out for about 8 months.
With world lab I will be able to create my school days(world) again..all my teen class mates.. football field..its gonna be like time travel!
Why did you go to London?
Cant use grok! There is no free tier in europe?
Hey Matt! While you're out there wrapping up your Spotify year, there’s one huge update you totally missed. The whole NotebookLM crew just left Google this week to kick off something new. Get ready, this could be the next big thing in AI tools! 🤖👀🚀 Stay tuned!
next time he goes for a world tour , AGI will drop
when you say X are you talking about twitter?
Nobody should be talking about either...😬
@@WilliamStonerockamen
Ur stuck in the past, good luck, adapting to AI
Wow 🙏🏿🙏🏿🙏🏿
Pretty soon there won't be enough electric power to run AI
Glad you specified the power as the 'electric' type
Already a problem.
We can build more.
They'll switch the hoomans off
🔥
Aesthetic, there is a 'th"
Where are timestamps?
It is in description but didn't know why it isn't showing
We need an ai for that 😂
I got a question: In your opinion on which year we reach Immortality and Longevity Escape Velocity?
My guess is immortality by 2045. Longevity escape velocity has already been happening for those who will live to see the day immortality becomes available and its definitely less than 75 years away.
@@SeverSFSs Understand.
200th comments : great explanation
Need time stamps
My code testing shows that the full o1 model (not pro) still sucks compared to Claude.
Oh man I just want Voice Mode. its all I want. Unlimited. I don't care for 01. I just want unlimited voice. FFS. "A few weeks they said".
So if Gemini is the top ranked AI should I still be paying for GPT? Gemini is free and GPT cost me 20 quid a month
Sam Tweeted congrats to David Sacks and Elon Responded “😂” lmao 2025 gonna be an interesting year indeed
Are you the guitar TH-camr who just cut his hair and pivoted to covering AI? You look and sound just like him.
Great Ai news. Open Ai Sora just seems to be all talk, and demos, while other video models have been running already. Amazon has also been sleeping and very slow to innovate via AWS. They just updated their terrible AWS console UIX. No major improvements compared to Google and Azure.
ginie2 is game changing for future everyone will be able to create his own game
05:00 War Games incoming..... * sarcasm *
TimeStamps Please!
Will A.I ever slow down????
No! Resistance is Futile! 😲
No.
Both Pandora's box and the proverbial floodgates are open. The cat is out of the bag and the toothpaste is out of the tube.
There's no going back. The future is here, the future is now.
Wolf wolf wolf
Meow
You can physically see that this video was completely generated with AI by how many cuts and edits are made during video playback.
Seems like the dead internet theory became reality.
plz provide us AI tool that we can use right now not only research
We're all living in a simulation. :)
From Coinbase ... Could be moving towards Agentic AI powered by crypto currency
ICP!
Flow State is a nice copy from certain nsfw ai pages
“Rapid prototyping” good wording. Id rather not play fully AI games.
Can someone make title time stamps please
It is in description but didn't know why it isn't showing
@@Brandswaggeray yeah right
an-doo-ril. That is how you pronounce Anduril
Wow! Millionaires cant even afford Chat GPT
the only update i want to hear is that these chatbots actually understand the words they are saying. until then like noam chomsky said 2 years ago about chatgpt and still remains true through now, is that these bot are just a parlor trick and bring us no closer to AI. As a computer scientist i could not agree more.
Arrangement of words indicates understanding of the meaning. Your variable is hidden in plain sight.
Grammar also dictates arrangement of words. Does grammar understand the meaning?
First comment. Big fan 🎉
Nah u aint
@ King, ain’t we all?
If we have human-sounding robots with artificial human-like organs then in 10 years there will for sure be real discussions of... uhmm... does a robot have to give consent.
Tron telemetric .... in the androidic consiones energetic constallations overlayerings levels of interaction, genuine computer accellerating towards the point were VR AR are being produced by genuine computer/s .... and not so much humans obtaining rightiones points for interaction in evolution creation...
the new marketing person had the great idea of the 200 dollars per month. Bye!
❤