This was a busy week in the world of AI.

I'll show you how you can now remove

watermarks from images and videos

generated by Gemini. There are expanded

features for notebooks. Chat GBT quietly

rolled out a really cool tool that for

some reason no one's talking about along

with some very dystopic features that I

don't know if even I feel comfortable

testing. I'll show you what those are

though along with all the other AI news

in this week's weekly AI news recap. My

name is Paul J. Lipsky and every week I

make a video just like this one where I

talk about the AI news that regular

people actually care about. So if that

interests you, make sure to subscribe.

And now let's get to the news. We'll

start with some really good news from

Gemini. Up until now, if you generated

any images or videos with Gemini, they

probably had this watermark on the

bottom right hand corner. Well, that's

changing right now. Gemini is rolling

out a new setting that lets you decide

whether your images, videos, and music

carry this watermark. So here, for

instance, on my free Gemini plan, I

asked it to create this image for me,

and there's no watermark. Now,

interestingly, Google has said that this

should be a setting that we can toggle

on or off. I don't see that anywhere.

It's just off for me, and it's not

creating any images with watermarks. And

I don't know who would turn it on, but

allegedly there's supposed to be a

setting somewhere at some point where if

you wanted to, you could turn it back

on. Importantly though, this does not

remove Google's invisible watermark,

which is called synth ID. Those will

remain embedded in the image, but the

visual watermarks will be removed. So,

if you download this image and then

upload it back into Gemini and ask, is

this AI generated or does this have a

synth ID? it will be able to detect that

and that will let you know that it was

generated or at least edited with AI.

So, that's not changing. But those

visual watermarks on the bottom right

hand corner, those are what's going

away. And this is a good thing. No one

else does this. And it's about time that

Gemini finally caught up. Now, if you're

a student, I also have some more good

news from Gemini. Starting this week,

students in the United States can claim

one year of Google AI Pro at no cost.

That's normally $20 a month and that

includes a lot of benefits. You get

Gemini, Gemini Spark, Gemini Inside

Gmail, 5 terabytes of storage and Google

Health Premium all for free. If you want

to grab that deal and you are a student,

you can just come over to this URL right

here 1.google.com/ai

student. And right here, you see it is

$0 a month. If you're not a student, I

wouldn't suggest you even try to grab

this deal because they have some sort of

verification to make sure you actually

are a student. If you're outside the

United States, you can claim instead one

year of Google AI Plus at no cost.

Google is also rolling out a new feature

called StudentHub, where Gemini will

collect all of its different learning

tools in one place to make it easier to

study for your classes. I don't know

exactly what this will include or how to

access this because I don't have it yet.

But what we do already have access to

are study notebooks. And most people

don't know about these even if they are

familiar with notebooks and notebook LM

or Gemini notebook as it's now called.

But here inside of Gemini over here on

the left under notebooks, these show you

all of your Gemini notebooks here inside

of Gemini. But if you click on new

notebook from within Gemini, you will

see an option here for a regular

notebook which is organize your ideas.

And they have this new one called study

and learn. And you can set up these

notebooks as a learning guide that will

quiz you and keep track of your

progress. Here, for instance, I created

one for an AP physics class. And if you

scroll down here, the first thing you

can do is personalize your learning by

taking a diagnostic quiz. This will kind

of get a baseline for how much you

already know. And then based on that, it

will give you these different study

guides. You can read through the

material. It gives you diagrams

sometimes. And then you can come down

here to take a quiz based on that

lesson. So you can make sure that you're

actually learning and retaining that

information. And then if you go back to

the notebook, you can get re-quized on

these topics over and over again to make

sure that they actually stick. I think

it's a very very useful tool for anyone

who's a student. I'm sure that's going

to make up a large part of that new

student hub. So yeah, if you're a

student, definitely take advantage of

the fact that this is going to be free

for you and it's going to include a ton

of resources to help you stay on top of

your studies. Before we get to the rest

of the AI news, let's take a quick break

to hear from today's sponsor, Code

Rabbit. These days, anyone can vibe code

an app using AI, but how do you actually

know if there are problems with it? As

someone who has never written a single

line of code manually in my life, I need

a tool that explains the code to me and

looks for any issues with it. That's

where Code Rabbit comes in. It's an AI

code review agent that looks at all the

changes that were made to the code,

summarizes them, and then suggests how I

can improve my app. Code Rabbit's new

change stack makes those changes easier

to follow. It groups files into logical

cohorts and dependencyaware layers

instead of listing them alphabetically

so I can understand what changed why and

how the pieces fit together. The

overview puts the summary blockers with

one-click fixes, reviews, and comments

in one place. Timeline shows how the

poll request reached its current state,

while semantic diff removes formatting

and moved code noise. Using Agentic

Chat, I can even ask it to explain the

code to me in a way that a non-coder

would understand. If you're like me and

you love to vibe code, but want another

set of eyes before you ship, then you

need to check out Code Rabbit. The link

for it is in the description down below.

And now back to the news. Okay, we're

not done with the news from Google yet.

Last week, I told you that Google had

released their newest model, Gemini 3.7

Flash. At the time though, it was only

available inside of certain tools like

Google Spark, but it is just rolled out

to more tools, including directly inside

of Gemini. So now inside of Gemini, as

long as you have a pro or ultra account,

you can select 3.7 Flash. 3.7 Flash also

now powers AI mode inside of Google

search and it's available inside of

Google Workspace tools like Sheets. And

that's consistent with the way Google

has been running things. They use Flash

to power all of their Google AI tools

across their ecosystem, not just inside

of Gemini. And since it's currently

their most powerful model, this actually

pretty nice. And you can be strategic

about this. As I'm sure you know, Gemini

has usage limits, but certain things

don't count against that usage limit,

including AI mode inside of Google

Search and Gemini Notebook. So, the more

that you can use those tools, the

further your Gemini subscription is

going to go. Speaking of Google search,

it looks like Google search is adding

notebooks, like the notebooks from

Notebook LM. So, check this out. Here's

Google search and you can see over here

on the left, he's opening up the left

side panel, and you'll see your

notebooks in here. similar to Gemini

where you can only see two notebooks

here and then you can add a new one but

this is inside of search. So when you

select a new notebook you can name it in

here. So this again looks very similar

to how you create notebooks inside of

Gemini. Select an icon for it. Add

sources. So yeah it just looks like the

same thing you can do in Gemini but

brought into search kind of unifying the

whole experience. I don't have this yet.

I'm very curious how this will work. I'm

assuming it syncs between all three,

Gemini Notebook, Gemini, and now search.

But I'm not really sure what this brings

to the table, how this is going to help

search, and how this is going to help

improve the notebooks that we already

have. Maybe it saves all the searches

inside the notebooks. I'm not quite

sure. I'm very curious and we'll

definitely be trying it when it comes

out. and I'll let you guys know what I

think of it. And the final announcement

from Google is that Gemini Live will now

be able to kick off deep research. I

love this. The more capabilities we can

bring to voice or live, the better in my

opinion. I use it for everything now as

much as I can because it just feels like

such a natural way to interact with AI.

So, Gemini Live is coming or Deep

Research is coming to Gemini Live. You

can ask it to look something up for you,

do uh some research for you. You can

then lock your phone or switch tasks and

it'll work in the background and you'll

get notified when it's done. Let's move

on now to OpenAI and talk about all the

updates to Chat GBT that are actually

important. There are two big

announcements I want to cover. First was

something that they quietly released

this past week. They didn't make a big

deal out of it at all, but I think it's

really cool and I think it deserves more

attention. So, check this out.

Currently, this is unfortunately only

available inside the Chat GPT web app.

But now, if you click on the left here

where it says library, you're going to

see a new section here called Google

Drive, assuming you've already connected

up your Google Drive inside of plugins.

But once you do that, you'll see Google

Drive here. You can click on it and you

can see all of the files you have inside

of Google Drive. Your docs, your sheets,

what else? PDFs that you have uploaded

in there. This is even a video I have in

there. And what's really neat is if you

click on any of these, it will open it

up inside of Chat GBT. Look over here on

the left. This is still chatbt, but

inside of Chat GBT is Google Drive. And

this has all the same capabilities of

Google Docs. So I can edit anything in

here. I can add new stuff. I can format

any of it. It's exactly the same way. It

works natively inside of Google Docs.

But because it's inside of Chat GBT, you

can ask Chat GBT to help you with this

document or whatever you have open by

coming down here to chat GBT. So, I

could just say, for instance, for this

example, I'll say, format this to make

it look nicer. Don't change any of the

text. That then opens up that request

inside a new chat. Okay. And I want to

keep this honest. We ran into an issue

here. This didn't happen when I tested

it before. But first, it got totally

stuck and I had to stop it and reprompt

it to try again. And now instead of

editing that document, which it didn't

when I tested it before, it created a

Word version of it, which is not what I

wanted at all. So, let's try this again.

We're going to try the other method,

though. I'll show you what I mean. I'm

going to come over here on the top left

and click on new chat. And this time,

I'm going to type in the at symbol and

just type in the name of the document.

This works too. If you type in the name

of the like Google doc for instance, it

will show up right here without you

having to go to the library section

first. So with that now typed in, I can

click on it and that will open it up in

this side panel on the right. So now I

can be chatting with chat GPT here while

working on the document on the right.

And let's try this again. I'll just tell

it to edit my Google doc. Don't change

any of the words, but I want you to make

the formatting look nicer here in the

doc. And there we go. Now, this probably

wasn't the best example because it was

already so well formatted, but this just

proves that, yeah, it can read what's

inside of Google Drive and make changes

to things like Docs. I don't know. I

think that's pretty cool. Instead of

having two windows open side by side,

one with Chat GBT and one with Google

Drive, or having to switch back and

forth, it can all be within Chat GBT. I

think that's pretty neat. I would love

to see this in the desktop app and for

more applications, maybe Notion or

somehow Obsidian on the desktop. That

would be awesome. The second

announcement from OpenAI that I want to

talk about is a feature they're calling

computer history. If you choose to turn

computer history on, then chat GBT will

create memories and a timeline of

everything that you do on your computer

to better help you with future tasks.

So, if we come over to the Chat GBT

desktop app and you come into the

settings over here on the left is a

setting for computer history. And if you

want, you can turn this on. As you see,

I've not turned mine on, and we'll talk

about that in a second. But once you

have it on, it will begin keeping a

memory timeline of things you're doing

on your computer. So, here's an example

of one. This user at 9:20 in the

morning, they reviewed their launch

plan. They checked the implementation

and they gathered feedback before

updating the documentation. That was the

memory that was recorded, and you can

see which apps were used for that. Then

at 9:10 they reviewed the project

feedback and they use Chrome chat GBT

and Slack for that. And down here they

organize their work for the day using

Chat GBT, VS Code and Terminal. So what

is the point of this? Well, here are

some examples. So let's say you ask Chat

GBT, what was I working on before my

last break? Chat GBT will normally say I

don't know. But with computer history on

it will check the computer history and

say oh you were revising a launch

presentation or if you ask it what was

that planning document I was looking at

today normally chat GBT will say I need

more detail but with computer history on

it will say you were reviewing the Q1

launch plan in Google Docs and here's

another example if you just let's say

uploaded a video to YouTube and you

asked Chat GBT or you told Chat GBT I

just uploaded ed my video, make a skill

for next time, it will say, "Well, let

me know what steps you followed." But

with computer history on, it will

actually be able to see what you did for

uploading that video and turn that into

a skill that it can then reuse in the

future. Now, I'm not 100% sure how

useful this actually is, and I'm not

going to be doing a demo for you because

this is something I don't feel

comfortable using yet. Obviously, I feel

comfortable using a lot of different AI

tools, but I do have a limit. I'm not

quite sure where that limit is, but this

one I feel like it does cross the line

for me. I just don't feel comfortable

with this yet. So, I'm not going to use

it. I'm probably being a little

overcautious here. It's probably fine.

But, I also don't see an overwhelming

need for this right now when all the

tools I already use work so well. And

it's pretty easy for me to just describe

to Chat GBT what I've been up to when I

need to catch it up on the work I'm

doing. I may change my mind on that in

the future, but for now, I'm leaving

this one toggled off. All right, we can

move on from OpenAI now and talk about a

couple of things that Anthropic has been

up to. First, they announced that Claude

Co-work is now available on mobile and

the web for all paid users. If you're

not caught up on this, originally Claude

Co-work was only available on the Claude

desktop app. However, recently they made

it more web-based. So, if I come to

claude.ai here, inside of Claude, I can

select co-work. And this was only

available on the more expensive plans at

first, but it has been rolling out and

is now available on all paid plans. Now,

Co-work in the cloud is a little bit

confusing because not all features are

available in the cloud compared to the

desktop app which can run locally.

Again, it can be a little bit confusing.

So, I've made an entire video on that

right here. I'll also remind you about

it at the end and link it up at the end

so you can figure out the difference

between cloud-based tasks and local

tasks with Claude Co-work. The second

announcement from Anthropic is that

Claude can now send emails in Gmail for

you and manage files in Google Drive.

And this is what that looks like here.

Now, why this has me scratching my head

is because I thought we could already do

this inside of Claude. For instance,

right here, I asked it to draft an email

for me. It drafted that email here. It

actually gave me three different

choices. And from this dropdown, I can

send this message via Gmail or I can

open it up inside of mail. But this

isn't the same UI as this. So, I don't

think I have the new roll out of this

feature yet, but yet I can still send

emails. So, this announcement made it

seem like Claude couldn't do this

before, and I'm I'm pretty sure it could

do that before. But anyway, just a heads

up that now the integration between

Claude and Gmail looks like it's going

to get a lot stronger, and that's also

going to include Google Drive. Let's end

the week with some rapidfire news

stories. So, first, have you guys seen

this video of the new AirPods that

leaked? We've known for a while that the

new AirPods are going to contain tiny

cameras in them, not for taking videos

or images, but for visual intelligence,

Apple's AI. This is the first time we've

seen them. I don't know how this video

got leaked, but it did. And in this

demo, the guy's looking at the book and

saying something like, "Remind me of

this later." the little cameras see the

book he's looking at and add that to his

reminders or whatever it's doing in the

background. So, this is the first time

we've seen these in action. Now, I am

very interested in the AI wearable space

and where it's headed, but the biggest

problem I see is adoption and comfort

for people to use AI wearables. It has

to come from a company that they trust

and implement it in a way that feels

like they can trust it. Having a camera

on glasses is not something that people

trust obviously, but inside of AirPods,

something that people already wear every

day and utilize in a way where you

cannot take photos or videos with it, it

can only be used for visual

intelligence. This may finally be a form

factor that people are willing to adopt.

Next, we have a bunch of quick stories

about Chat GPT. So, first, Chat GBT

released Chat GBT teens. This is

specifically built for learning and

there's a lot of protections built in

like if they try to go off track from

their studies, it will bring them back

to their studies. There's a lot of

parental controls. So, if you have a

teen and you want them to be able to use

AI tools and to learn how to use them,

check out ChatGBT teens. OpenAI has also

released an Apple messages plugin for

the Chat GBT desktop app. This will

allow Chat GBT to search messages, catch

up on conversations, draft and send

replies all from within Chat GBT. I

think there are a lot of useful use

cases and capabilities that this

unlocks. However, once again, this is

one of those areas where I don't feel

comfortable with this. I have a lot of

private messages on iMessage and I don't

want an AI looking through all those.

So, this is another one that I am not

going to be using. OpenAI has also

announced that you can add teammates as

editors to your chat GBT sites. Sites is

a way that you can quickly build

websites using chat GBT. I've made an

entire video about it before. I'll have

that linked in the description down

below if you want to check it out and

all the things you can create with Chat

GPT sites because I found it to be

incredibly useful. But now you can

invite a teammate to edit the site as

well. And finally, we'll end with some

slightly humorous news, or I at least

thought it was funny. There's this guy

representing himself in a Connecticut

court, and he made an official filing

with the court, and inside of it, he hid

a prompt injection. So, written in tiny

white letters that humans wouldn't be

able to see were instructions telling

any AI that was reviewing the document

to side with him in the case. Obviously,

this didn't work for a number of

reasons. One being that a court isn't

going to be relying on AI for writing

its decisions, or at least I hope not.

So, obviously, this failed, but I don't

know. I thought this was pretty funny

and pretty bold that he actually tried

that. So, there we go. That's all the AI

news from the past week that I thought

actually mattered. If you enjoyed this

and you want to make sure that every

week you get notified about my newest

weekly AI news recap where I cover

topics just like this, make sure to

subscribe to the channel and turn on the

bell notification. In the meantime

though, you can check out this video

right here. This is the one I mentioned

before where I talk about Claude Co-work

and how it works in the cloud versus

locally. So, if you're a little bit

confused about that or just want to

learn more about Claude Co-work, check

out this video right here and I'll see

you over there in 1 second. Bye for now.