AI News: Things Just Got Absolutely Crazy!
Вставка
- Опубліковано 11 чер 2024
- Here's the AI News that you might have missed. Subscribe to this channel and the newsletter for a chance to win the Insta360 X4 (winner announced May 31st).
Discover More From Me:
🛠️ Explore thousands of AI Tools: futuretools.io/
📰 Weekly Newsletter: www.futuretools.io/newsletter
🎙️ The Next Wave Podcast: @TheNextWavePod
😊 Discord Community: futuretools.io/discord
❌ Follow me on X: x.com/mreflow
🧵 Follow me on Instagram: / mr.eflow
Resources From Today's Video:
Introducing GPT-4o - UA-cam: • Introducing GPT-4o
Brockman on GPT-4o Image Capabilities - X: / 1790869434174746805
Hello GPT-4o | OpenAI: openai.com/index/hello-gpt-4o
Altman on GPT-4o and AI Future - UA-cam: • Sam Altman talks GPT-4...
Sutskever Leaves OpenAI - X: / 1790517455628198322
Altman on OpenAI Leadership - X: / 1790518031640347056
OpenAI Staffers Leaving: gizmodo.com/openai-staffers-r...
Google I/O 2024 Announcements: blog.google/technology/ai/goo...
JScigala on Google I/O - X: / 1790536121497399613
Anthropic Hires Instagram Co-founder: techcrunch.com/2024/05/15/ant...
AnthropicAI Announcement - X: / 1788958483565732213
Anthropic Console: console.anthropic.com
Chatter • Hume AI: chatter.hume.ai
Tsarnick on AI Discussion - X: / 1788826317712245107
Meta's AI Earphones - The Information: www.theinformation.com/articl...
Microsoft Build 2024 Preview: www.engadget.com/what-to-expe...
Sponsorship/Media Inquiries: tally.so/r/nrBVlp
Mailing Address: 3755 Avocado Blvd Unit 287, La Mesa, CA 91941
#AINews #AITools #ArtificialIntelligence
Time Stamps:
0:00 Intro
0:29 OpenAI Announcements
6:04 ChatGPT Features They DIDN'T Announce
10:42 Ilya (and others) leave OpenAI
14:43 Google's 100 Announcements!?!
17:15 Google's New Glasses
18:39 More Google Announcements
26:00 Anthropic Hires Instagram Co-Founder
26:30 Claude's Prompt Builder
27:41 AI Interactive News Radio
29:20 The Future of Dating with AI
30:17 Meta's AI Earphones
31:02 Microsoft Build
33:29 Contest And Final Notes - Наука та технологія
I'm one of the a lot of people who just watch the end of the week recap every week
Yes a weekly is needed
Same. I'm subscribed to too many AI channels to watch everything 😂
stop that
I mean if you want just do one long video a week :) unless the little vids are making you a lot then keep it up
Honestly.. I wish I was man. Lol
Oh man, WHEN google intergrates its AI into glasses, my entire world will be changed.
This is literally why I keep upgrading my pixel.
I was about to pre order focals by north when Google bought them never to be heard from again.
I definitely remember that episode of Black Mirror!
Yeah! Then next will come self-cleaning Google contact lenses that you won’t be able to take off! 😄 yay society!
I KNOW! I MIGHT FINALLY KNOW WHAT IT'S LIKE TO HAV A GIRLFRIEND! 💘
Access to the GPT store is pointless as no Custom GPTs are available for free, you need the paid plan to use any of them.
I legit just had a verbal conversation with GPT 4o about a very personal situation (that really needs a counselor/therapist and she walked me through the situation, I even got emotional. It was like instant therapy. Very thoughtful and helpful. She helped my write a letter about my feelings, that I can send or keep for myself. I seriously feel weight off my shoulders.
the new voice model they demo'd isn't available yet, is it? You can only use the old one from regular gpt4.
Voice model not out yet ...?
How did you do that when the new interface has not been released (only the model)?
I opened chat gpt select gpt4o and click on the headset icon to the right side of the chat box
@@gd8205 iOS, Android, Web? I'm looking forward to trying it. Both Google and ChatGPT pushed a lot of vaporware this week
I don't like how bubbly the chatgpt is, I just want an old British man as ai assistant.
lol. Like David Attenborough?
@@josephinemurphy2502exactly
How hideous. Lots of better elderly English voices than that.
Haha are you Batman
I'm going to have it wake me up in an annoying voice. Until I regret it/
I can see something like those google glasses becoming a headset for the blind, watching everything that's going on around them and answering questions they put to a mouth-mic.
The Open Ai presentation shows why smart Glasses have a future. Camera in the smartglasses can see what the user sees and comment and help to indentify and solve problems you see. And with bone conduction silently speak to the user
Given the backlash there was against Google Glass ("glass holes") I have my doubts. People don't like being filmed without their knowledge (even if, in most places in the world, it is perfectly legal to do so in a public place).
@@cbnewham5633they can buy a device that distorts there face if they really don’t like it.
@@cbnewham5633 Ha! To bad.
@@cbnewham5633 They are going to very quickly get used to it. It happens anyway with livestreaming, and then we have glasses like the Meta Ray-Bans, and let's not forget dash cams.
@@cbnewham5633tell that to Meta and RayBan. There are billions of mobile phones with up to 4 cameras on them
What the hell is a Car-Rick-a-chair? 😂😂
😂😂😂😂
It's a care-rick-cheer. 🤣
Some weird pirate way of saying it.
A rickety old chair
I had to come here and post that. That was wild. And he said it a few times 😂
Cheers for the update. Really appreciate it. No way I could keep up with developments without your channel! Ian, UK
Chatter is wild! Imagine all those AI news consolidation websites except now you can talk with the person delivering the news to you and getting a custom feed of exactly the news you care about!
I love your insight, around 25:00, that there is humanity involved. Real people who are genuinely trying to add value. Thank you 😊
I also loved that, especially because I had agreed more with the X post when I usually agree with Matt. But as soon as he started explaining, I once again confirmed how much I love his way of thinking.
I’m so glad I found this channel. You do an excellent job of keeping us informed and excited about AI. Thank you!
Imagine instead of a dating concierge the app helps you find friends in your area! It's so hard these days for adults to find and build communities. Something like this could help with the massive loneliness epidemic. Sure some people will want to be friends with their personal AI, but having AI find real friends for you will be game changing!
So... A church directory for your area?
😅
That's what Gorilla Guides will be, your happy, immersive, AI friends and passion concierge which Binary Bridge the Guerilla Nexus with the World !
I suspect AI will soon replace the need for human friends.
@@SuperheroArmorychannel That can't ever be replaced. Even with robotics lmao. Humans evolved to interact with each other, removing that social relationship is a huge contributor to mental health issues
@@coletcyre Never say never.
Regarding glasses and earbuds, imagine going to France and not speaking French and you want to speak to someone who doesn't speak English. You both have active AI earbuds. They speak French and you hear English. You speak English and they hear French. If you've got the glasses, you look at a French newspaper and it's translated into English for you as you're reading it. Now imagine you're on a road trip and you want to drive somewhere but you don't know the way. So you put your glasses on and through the glasses you see a line on the road, showing you the route as you drive without you having to look down at the sat nav.
It's this kind of every day use stuff that will be the key to widespread adoption in the future if these technologies can be made to fit into small products with healthy battery lives.
Imagine people form relationships and pair bonds enabled by this tech then a few years in its all lost due to a solar storm say. Babylon.
@@kristinabliss I'd read this novel!
AI Interactive news radio is crazy.
The end of Matt’s channel ?
I think I know why they are offering so much stuff for free - they need content generators. We Free Users are giving them content to let their models learn more, all at no charge. It's a good deal.
Also they need AI to become mainstream and accepted by all. Even the skeptics.
I remember when the internet was considered a fad
Well I’m in India , and I had access to gpt 4o , including the voice ( which btw pi Ai does wonderfully already) I used it to half an hour till it rolled back to 4o with limited capacity
They have to compete with Google now. They need to do something to sweeten the deal.
@@gd8205 Not true, Internet was accepted when the ease of use and applications of it were there.
AI will be accepted when its easy to use and the use cases are there all technology goes through those phases.
Yep, its a data collecting exercise.
Something no one talks about, is the fact that as a Chat GPT Plus subscriber in Europe, you do NOT get the memory feature. But you get it in the free tier in the US... kinda pisses me off!!
Hey Matt, thanks for the week summary and going to Google IO and reporting back to us. Nice shirt. :) Love the positive energy too.
I totally missed the glasses part, good catch!
It's exciting to see all the developments every week. Thanks Matt!
Matt, thank you for sharing your enthusiasm with us. Wish I could do the same, but also know that this takes so much more time than I can afford.
Thank you for being the best news source on the internet. Now I am very excited to settle in, watch this and then spend the weekend with AI lol
All of the openai demo videos are directly on their youtube channels. I think they are better than the main examples on the stage.
Thank you for consistently bringing so much "Signal' to your video reports, Matt!!
Thank you for your content. This channel is my favorite for your focused content!
Curious. After AI creates a logo for a specific product. Can you get it to give you the code for 3D printing the product in CAD and the laser code and the best material?
even though i watch all same the conferences @ 2x / 4x, i still come here for your takes and observations
thanks Matt
Matt, I've followed your channel from the start. I love it. You are my new "Star Trek".
Thanks for the great news from Germany. Can't wait for AGI robots replacing me on my boring job, while I'm spending the hopefully upcoming universal basic income with my gf going to the cinema or a city trip...😮 (GDP is generated by robots then).
Thanks as always Matt. It's so interesting that we're in this phase of AI intergrating with daily life via the way we emails, the way we date, our productivity etc. It goes to show we aren't just going to wake up to a changed world. It's happening very progressively by changing the infrastructure of our lives and how we do everyday things.
What image generation tool can I use for consistent story telling ?
Read the T&Cs! Everything you input to GPT-4o is, by default, used to train their models. You can opt-out, otherwise you're giving them your data, your pictures (and likenesses within them) and your actual voice.
Thanks for this summary. Well presented and well paced. Learned a few new things which is what I'm always hoping. Keep it up!
4o is good and I have not messed around with the new image and sound generation yet. Still not totally game changing for daily work though. Not a personal assistant yet. I need it to manage my calendar and remind me about stuff. An alarm I can just talk to and tell it when to wake me up or have it tell me I need to leave for a meeting. Also, the environment in which it runs code is super restrictive. I think that is more a barrier to making graphs and diagrams than its ability to code. Finally, it needs photo memory, not just text memory, so it can remember people it is looking at. Once it has all that, I think it is basically AGI.
Next week prob
The images are wild
Honestly I agree with Joshua 24:18, I also want to add that sometimes they don’t even release the features they tease (or it takes several years before it comes, and at that point is behind). It’s like the boy that screamed wolf
lmao, the way you timestamped Joshua made it look like a Bible verse.
The rabbit R1 got an update can you give us some updates how it is running from time to time
Thank you, Matt, for this great summary. This is so fascinating!
Back then, when they released a keynote or a new update, you could use it right away. But after Google made Gemini multimodal, they showcased the so-called update, but it wasn't actually available yet. Worse, it didn't even have the capabilities they claimed. Now, the succeeding update release keynotes just showcase what might be possible in the future or ideas they hope to implement. Why not create or finalize it first, then release the keynotes or updates when it's ready to use immediately
Hey Matt! If you take off the chair headrest, it won’t be flickering madly when you do the ‘green screen.’ It looks like you’re being attacked by a drunken bat and is very distracting. Easy fix!
I can't access gpt 4o for free? Is that a country restriction?
The rollout to free will happen over several weeks. So you just have to be a bit patient.
Just a total package. I think GPT-4o is the beginning of what we have all been looking for. It's only going to get better.
When it can talk with me while cleaning my dishes and organizing them I will buy.
I agree its very smooth to use, I just wish they would have let us build GPTs with it, now that would be super useful. It does do some other very useful things though.
That's a dangerous fallacy. But there is still hope.
always love your content and excitement!!
nice rundown, it's hard to keep up - I am a bit lost on where I can actually talk and interrupt 40 on my mobile? I have an app, and can talk to it but it doesn't want to converse with me. No surprises there though :) My phone is 'old - an 128gb X' so that might be it. Thanks for your succinct precis' ...
Anthropic Prompt Generator was not available when I checked....wonder if its for specific partners right now....the whole button was missing when I logged in....and I pay for Claude Opus.
Thank you so much for your explanation, I've come back to watch a second time! This is mind blowing!
just love what you're shared, this week AI news is Tsunami, information overload.
Really appreciate weekly updates as a fellow tech enthusiast
I really just wanna talk about how it helped a blind person navigate and described the world as he was walking around. As a completely blind person myself, I’m extremely excited for this.
That chatter news was awesome, a very interesting use case
So as Ilya and Jan understood, we're at the edge of the abyss.. And this week, we took a brave step forward 🙂.
All these improvements and updates are really beginning to compound, accumulate, and accelerate. The next few years are going to be something special to see
23:17 the perfect tool for training conversational speech
Thank you for keeping us up to speed.. It is wild!
Spot on about the glasses. I have the Ray Bans as well as a cheap chinese pair that work fine for anything but high fidelity, and you see the tell tale large arms needed to hold the battery and processor, etc.
The Ray Bans are awesome, wouldn't surprise me if Google needs to get back in this game as I feel these are close to mass market and will fall under $200 in a year likely from competition.
I don't see them having a display though, there's always tell tale signs you can see in the lens for them as well as likely some extra space for the extra display circuitry. I think that was just a text overlay afterwards.
Good catch on the blog Matt! That's why you are good, you put the work in.
4o free has big problems. If you are outside the US, it starts to block you after two additional asks. Also the free tier has too few requests to work with it productively (for example during a programming session), so its fun to play around with, but ultimately pretty useless if you have to use it productively... That would be ok (they need to get paid), if they hadn't announced its free for everyone implying its really usable...
Even the paid subscription for GPT-4 isn't really usable at this point due to the limited number of requests.
Hmm, very impressive - and huge pivacy risks. If the phone listens in on all my phonecalls. How do I opt out?
Did Hume say “ruge waves”? 28:49
Ooooh bonus giveaways! Love that for us :)
New camera or just new settings? 24 fps and 2k resolution is cool (the display on the background probably not in sync with these )
Continue with this great content!
Matt's channel is best so far for getting up to date with AI news, is their any other such channel available?
I checked this morning, still no 4o, great sadness.😨
Paid or free? I'm in the UK and both have 4o as of yesterday.
I have that, but the GPT Store doesn't seem accessible to the genera public, yet.
I have 4o, but the features they demonstrated aren't yet available.
@@m0ose0909 Same
Great stuff keeping on top of so much stuff this week.
I downloaded the new ChatGPT Mac app, but don't see the option to allow it to see what's on my screen, other than uploading a screenshot file. Maybe I'm not looking in the right place, or has not it not been released yet?
I've got access to GPT-4omni, but it absolutely refuses to output any image. I read that the free model supposedly has this feature, but if so, I certainly don't have access to it.
@everyone is there already a chance to use the videofunction or is it about to be released in the future?
Matt, you're simply doing too good of a job. Take 6 months off. Relax. Unwind on a pink sandy beach in Bermuda. Enjoy some rum swizzles with your toes in the water. Play a round of golf, or two, at Port Royal and Mid Ocean. TAKE IT EASY. By the time you come back, most of us here will just be starting to catch up on ALL your videos posted to date. There's just too much to stay on top of. So you can relax as we on the other hand have a lot of notes to take and material to digest.
You've earned it. Thanks 😊
GPT-4o _does_ limit your use at the moment and it's not as good as most say (atm); I uploaded a picture of a well known landmark and it misidentified it. Love these teach-in videos of Wolfey's though, they keep me in the loop in this everchanging landscape.
GPT-4o hasn't been fully released yet. Most features are missing.
I believe I understand why they didn't invest in showcasing the "exploration of capabilities" examples in the live demo; these features are not functioning yet, even for users with version 4o.
Just a guess. But unlike most windows users complaining, I assume it wasn’t OpenAIs choice alone to first release the Mac desktop app. Microsoft wants Copilot to become the default windows assistant. OpenAI has to wait until Microsoft decides copilot has a strong position or moat
The reason open AI didn’t show off the other capabilities is because they do not work nearly as good as they say/show they do. I just tried all of them and not once did I get the result they claimed I would. Some were a bit better than others but again, none performed the way open AI claimed they would.
Chat GPT-4o hasn't even been fully released to the public yet? You're a liar.
They also said these functions were going to be released over the next few weeks, so while you may have 4o (like me), you may not have access to the new features yet (also like new).
@@CVRogers174o has been released, just not so the features they showed.
I have it, just not the features.
Fr@@CVRogers17
Funny to hear people reacting to the voice model that had been around for a year thinking it is the new model.
"able to ask your phone a question and get an answer" what i wanted my phone to do since voice assistants came out. Which took years if not a decade to be able to turn on the flashlight; beyond asking what the weather is.
Do you have to be a paid member for The Prompt Generator ?
Just found this channel, you do a great job!
I always have trouble with discord. I click link to join, accept invite, it takes me to the playstore and have to click through to open app, as I have it already, then nothing. Discord opens and no sign of channel. Android phone? Any suggestions ?
You need iPhone or Mac to use a lot of the features?
That is a google glass I'd be VERY interested in buying!
That comparison between openai and google and where to go to use the stuff is spot on. Also, google just randomly cancels stuff, so you should be weary trying to build stuff with their new tech.
5:25
Me: Gpt, what's the weather like this week?
*INITIATING DRAMATIC ROBOTIC VOICE*
Me: 👀....fuck
Bot body with gpt to fuck not quite out yet.
So I haven't been able to generate an image for several days in gpt. I just tried again in 40 and it tells me sorry I'm not capable of generating images. I don't remember this being an issue in the past. Did they turn that feature off?
I had that happen randomly a few months ago. It was even saying it right before or after it complied and produced an image. There is a whole lot of things going on inside the black box of OpenAI methinks
now when can WE get to try it out like the demos... still waiting :)
Do you use midjourneys character ref for your thumbnails or another tool.? The likeness is awesome
would love google glass 2.o
Nothing is more impressive than that WONDERFUL human-like "HER" voice. FREAKING AMAZING!🥰
Yay. Been checking every day. I know have 4o. Love your work by the way. Legend. 👍
The voice speed in the video is cut. Im using gpt4-o right now and its not even close to that responsive...i live in sweden. Someone know why?
Awesome Matt. Count me in.
Don't see "Generate a prompt" in that popup... Maybe limited access?
Where exactly can I use this new Chat GPT model?
Looking fire in your thumbnail Matt lol
Apple Vision Pods would be dope. Cameras on AirPods would be sick to just ask about things as you are around without having things in your eyes. I’ve thought about this for a while since that one girl from Apple mentioned that sensors are going to be a huge area of interest with AirPods.
Google glasses with "total recall" could also help prevent a lot of crime, especially with some sort of feature that automatically upload potential crimes in progress to the police. I would be curious if it could identify all the different plants and their uses while out hiking.
I went to GPT and it asked if I want to use 4o. I agreed and then asked it to generate an image but it only responded with the standard text based answers - what am I missing?
i started using 4o for coding and terminal commands instead of gpt4 snd 4o performed worse, switched up command words eg apply instead of append
finally a good use for a smartwatch or smart glasses, goes perfectly with gpt-4o
This is probably one of the most promising updates so far. Hang on people, it's a great time to be alive!
Excellent video. Minor note, your stress for the word "caricature" is on the wrong syllable. Your stressing the 2nd and it's supposed to be the 3rd.
Based on what you say this week Matt. Which option would you opt for? Cheers David
I think it was a bit suspicious that Mira got a tweet asking if it could do real time translation, when they'd just posted a pre-recorded video on their website of it doing exactly that 🤔
thank you