Google’s Gemini AI can now handle bigger prompts thanks to next-gen upgrade

Google’s Gemini AI has only been around for two months at the time of this writing, and already, the company is launching its next-generation model dubbed Gemini 1.5.

The announcement post gets into the nitty-gritty explaining all the AI’s improvements in detail. It’s all rather technical, but the main takeaway is that Gemini 1.5 will deliver “dramatically enhanced performance.” This was accomplished with the implementation of a “Mixture-of-Experts architecture” (or MoE for short) which sees multiple AI models working together in unison. Implementing this structure made Gemini easier to train as well as faster at learning complicated tasks than before.

There are plans to roll out the upgrade to all three major versions of the AI, but the only one being released today for early testing is Gemini 1.5 Pro. 

What’s unique about it is the model has “a context window of up to 1 million tokens”. Tokens, as they relate to generative AI, are the smallest pieces of data LLMs (large language models) use “to process and generate text.” Bigger context windows allow the AI to handle more information at once. And a million tokens is huge, far exceeding what GPT-4 Turbo can do. OpenAI’s engine, for the sake of comparison, has a context window cap of 128,000 tokens. 

Gemini Pro in action

With all these numbers being thrown, the question is what does Gemini 1.5 Pro look like in action? Google made several videos showcasing the AI’s abilities. Admittedly, it’s pretty interesting stuff as they reveal how the upgraded model can analyze and summarize large amounts of text according to a prompt. 

In one example, they gave Gemini 1.5 Pro the over 400-page transcript of the Apollo 11 moon mission. It showed the AI could “understand, reason about, and identify” certain details in the document. The prompter asks the AI to locate “comedic moments” during the mission. After 30 seconds, Gemini 1.5 Pro managed to find a few jokes that the astronauts cracked while in space, including who told it and explained any references made.

These analysis skills can be used for other modalities. In another demo, the dev team gave the AI a 44-minute Buster Keaton movie. They uploaded a rough sketch of a gushing water tower and then asked for the timestamp of a scene involving a water tower. Sure enough, it found the exact part ten minutes into the film. Keep in mind this was done without any explanation about the drawing itself or any other text besides the question. Gemini 1.5 Pro understood it was a water tower without extra help.

Experimental tech

The model is not available to the general public at the moment. Currently, it’s being offered as an early preview to “developers and enterprise customers” through Google’s AI Studio and Vertex AI platforms for free. The company is warning testers they may experience long latency times since it is still experimental. There are plans, however, to improve speeds down the line.

We reached out to Google asking for information on when people can expect the launch of Gemini 1.5 and Gemini 1.5 Ultra plus the wider release of these next-gen AI models. This story will be updated at a later time. Until then, check out TechRadar's roundup of the best AI content generators for 2024.

You might also like

TechRadar – All the latest technology news

Read More

From chatterbox to archive: Google’s Gemini chatbot will hold on to your conversations for years

If you were thinking of sharing your deepest, darkest secrets with Google's freshly-rebranded family of generative AI apps, Gemini, just keep in mind that someone else might also see them. Google has made this explicitly clear in a lengthy Gemini support document where it elaborates on its data collection practices for Gemini chatbot apps across platforms like Android and iOS, as well as directly in-browser.

Google explained that it’s standard practice for human annotators to read, label, and process conversations that users have with Gemini. This information and data are used to improve Gemini to make it perform better in future conversations with users. It does clarify that conversations are “disconnected” from specific Google accounts before being seen by reviewers, but also that they’re stored for up to three years, with “related data” like user devices and languages as well as location. According to TechCrunch, Google doesn’t make it clear if these are in-house annotators or outsourced from elsewhere. 

If you’re feeling some discomfort about relinquishing this sort of data to be able to use Gemini, Google will give users some control over how and which Gemini-related data is retained. You can turn off Gemini App Activity in the My Activity dashboard (which is turned on by default). Turning off this setting will stop Gemini from saving conversations in the long term, starting when you disable this setting. 

However, even if you do this, Google will save conversations associated with your account for up to 72 hours. You can also go in and delete individual prompts and conversations in the Gemini Apps Activity screen (although again, it’s unclear if this fully scrubs them from Google's records). 

A direct warning that's worth heeding

Google puts the following in bold for this reason – your conversations with Gemini are not just your own:

Please don’t enter confidential information in your conversations or any data you wouldn’t want a reviewer to see or Google to use to improve our products, services, and machine-learning technologies.

Google’s AI policies regarding data collection and retention are in line with its AI competitors like OpenAI. OpenAI’s policy for the standard, free tier of ChatGPT is to save all conversations for 30 days unless a user is subscribed to the enterprise-tier plan and chooses a custom data retention policy.

Google and its competitors are navigating what is one of the most contentious aspects of generative AI – the issues raised and the necessity of user data that comes with the nature of developing and training AI models. So far, it’s been something of a Wild West when it comes to the ethics, morals, and legality of AI. 

That said, some governments and regulators have started to take notice, for example, the FTC in the US and the Italian Data Protection Authority. Now’s a good time as ever for tech organizations and generative AI makers to pay attention and be proactive. We know they already do this when it comes to their corporate-orientated, paid customer models as those AI products very explicitly don’t retain data. Right now, tech companies don’t feel they need to do this for free individual users (or to at least give them the option to opt-out), so until they do, they’ll probably continue to scoop up all of the conversational data they can.


TechRadar – All the latest technology news

Read More

Google’s new generative AI aims to help you get those creative juices following

It’s a big day for Google AI as the tech giant has launched a new image-generation engine aimed at fostering people’s creativity.

The tool is called ImageFX and it runs on Imagen 2,  Google's “latest text-to-image model” that Google claims can deliver the company's “highest-quality images yet.” Like so many other generative AIs before, it generates content by having users enter a command into the text box. What’s unique about the engine is it comes with “Expressive Chips” which are dropdown menus over keywords allowing you to quickly alter content with adjacent ideas. For example, ImageFX gave us a sample prompt of a dress carved out of deadwood complete with foliage. After it made a series of pictures, the AI offered the opportunity to change certain aspects; turning a beautiful forest-inspired dress into an ugly shirt made out of plastic and flowers. 

Image 1 of 2

ImageFX - generated dress

(Image credit: Future)
Image 2 of 2

ImageFX - generated shirt

(Image credit: Future)

Options in the Expressive Chips don’t change. They remain fixed to the initial prompt although you can add more to the list by selecting the tags down at the bottom. There doesn’t appear to be a way to remove tags. Users will have to click the Start Over button to begin anew. If the AI manages to create something you enjoy, it can be downloaded or shared on social media.

Be creative

This obviously isn’t the first time Google has released a text-to-image generative AI. In fact, Bard just received the same ability. The main difference with ImageFX is, again, its encouragement of creativity. The clips can help spark inspiration by giving you ideas of how to direct the engine; ideas that you may never have thought of. Bard’s feature, on the other hand, offers little to no guidance. Because it's less user-friendly, directing Bard's image generation will be trickier.

ImageFX is free to use on Google’s AI Test Kitchen. Do keep in mind it’s still a work in progress. Upon visiting the page for the first time, you’ll be met with a warning message telling you the AI “may display inaccurate info”, and in some cases, offensive content. If this happens to you, the company asks that you report it to them by clicking the flag icon. 

Also, Google wants people to keep things clean. They link to their Generative AI Prohibited Use Policy in the warning listing out what you can’t do with ImageFX.

AI updates

In addition to ImageFX, Google made several updates to past experimental AIs. 

MusicFX, the brand’s text-to-music engine, now allows users to generate songs up to 70 seconds in length as well as alter their speed. The tool even received Expressive Chips, helping people get those creative juices flowing. MusicFX even got a performance boost enabling it to pump content faster than before. TextFX, on the other hand, didn’t see a major upgrade or new features. Google mainly updated the website so it’s more navigable.

MusicFX's new layout

(Image credit: Future)

Everything you see here is available to users in the US, New Zealand, Kenya, and Australia. No word on if the AI will roll out elsewhere, although we did ask. This story will be updated at a later time.

Until then, check out TechRadar's roundup of the best AI art generators for 2024 where we compare them to each other. There's no clear winner, but they do have their specialties. 

You might also like

TechRadar – All the latest technology news

Read More

Microsoft Edge could soon get its own version of Google’s Circle to Search feature

As the old saying goes, “Imitation is the sincerest form of flattery”. Microsoft is seemingly giving Google a huge compliment as new info reveals the tech giant is working on its own version of Circle to Search for Edge.

If you’re not familiar, Circle to Search is a recently released AI-powered feature on the Pixel 8 and Galaxy S24 series of phones. It allows people to circle objects on their mobile devices to quickly look them up on Google Search. Microsoft’s rendition functions similarly. According to the news site Windows Report, it’s called Circle To Copilot. The way it works you circle an on-screen object with the cursor – in this case, it’s an image of the Galaxy S24 Ultra

Immediately after, Copilot appears from the right side with the circled image attached as a screenshot in an input box. You then ask the AI assistant what the object is in the picture, and after a few seconds, it’ll generate a response. The publication goes on to state the tool also works with text. To highlight a line, you will also need to draw a circle around the words.

Windows Report states Circle To Copilot is currently available on the latest version of Microsoft Edge Canary which is an experimental build of the browser. It’s meant for users or developers who want early access to potential features. The publication has a series of instructions explaining how you can activate Circle To Copilot. You'll need to enter a specific command into the browser's Properties menu.

If the command works for you, Circle To Copilot can be enabled by going to the Mouse Gesture section of Edge’s Settings menu and then clicking the toggle switch. It’s the fourth entry from the top.

Work in progress

We followed Windows Report's steps ourselves; however, we were unable to try out the feature. All we got was an error message stating the command to activate the tool was not valid. It seems not everyone who installs Edge Canary will gain access, although this isn’t surprising. 

The dev browser is, not surprisingly, unstable. It’s a testing ground for Microsoft so things don’t always work as well as they should; if at all. It is possible Circle To Copilot will function better in a future patch, however, we don’t know when that will be rolling out. We are disappointed the feature was inaccessible on our PC because we had a couple of questions. Is this something that needs to be manually triggered on Copilot? Or will it function like Ask Copilot where you highlight a piece of content, right-click it, and select the correct option in the context menu?

Out of curiosity, we installed Edge Canary on our Android phone to see if it had the update. As it turns out, no. It may be Circle To Copilot is exclusive to Edge on desktop, but this could change in the future.

Be sure to check TechRadar's list of the best AI-powered virtual assistant for 2024.

You might also like

TechRadar – All the latest technology news

Read More

Google’s impressive Lumiere shows us the future of making short-form AI videos

Google is taking another crack at text-to-video generation with Lumiere, a new AI model capable of creating surprisingly high-quality content. 

The tech giant has certainly come a long way from the days of Imagen Video. Subjects in Lumiere videos are no longer these nightmarish creatures with melting faces. Now things look much more realistic. Sea turtles look like sea turtles, fur on animals has the right texture, and people in AI clips have genuine smiles (for the most part). What’s more, there's very little of the weird jerky movement seen in other text-to-video generative AIs. Motion is largely smooth as butter. Inbar Mosseri, Research Team Lead at Google Research, published a video on her YouTube channel demonstrating Lumiere’s capabilities. 

Google put a lot of work into making Lumiere’s content appear as lifelike as possible. The dev team accomplished this by implementing something called Space-Time U-Net architecture (STUNet). The technology behind STUNet is pretty complex. But as Ars Technica explains, it allows Lumiere to understand where objects are in a video, how they move and change and renders these actions at the same time resulting in a smooth-flowing creation. 

This runs contrary to other generative platforms that first establish keyframes in clips and then fill in the gaps afterward. Doing so results in the jerky movement the tech is known for.

Well equipped

In addition to text-to-video generation, Lumiere has numerous features in its toolkit including support for multimodality. 

Users will be able to upload source images or videos to the AI so it can edit them according to their specifications. For example, you can upload an image of Girl with a Pearl Earring by Johannes Vermeer and turn it into a short clip where she smiles instead of blankly staring. Lumiere also has an ability called Cinemagraph which can animate highlighted portions of pictures.

Google demonstrates this by selecting a butterfly sitting on a flower. Thanks to the AI, the output video has the butterfly flapping its wings while the flowers around it remain stationary. 

Things become particularly impressive when it comes to video. Video Inpainting, another feature, functions similarly to Cinemagraph in that the AI can edit portions of clips. A woman’s patterned green dress can be turned into shiny gold or black. Lumiere goes one step further by offering Video Stylization for altering video subjects. A regular car driving down the road can be turned into a vehicle made entirely out of wood or Lego bricks.

Still in the works

It’s unknown if there are plans to launch Lumiere to the public or if Google intends to implement it as a new service. 

We could perhaps see the AI show up on a future Pixel phone as the evolution of Magic Editor. If you’re not familiar with it, Magic Editor utilizes “AI processing [to] intelligently” change spaces or objects in photographs on the Pixel 8. Video Inpainting, to us, seems like a natural progression for the tech.

For now, it looks like the team is going to keep it behind closed doors. As impressive as this AI may be, it still has its issues. Jerky animations are present. In other cases, subjects have limbs warping into mush. If you want to know more, Google’s research paper on Lumiere can be found on Cornell University’s arXiv website. Be warned: it's a dense read.

And be sure to check out TechRadar's roundup of the best AI art generators for 2024.

You might also like

TechRadar – All the latest technology news

Read More

Here’s your first look at Google’s new AI Assistant with Bard, but you’ll have to wait longer for a release date

2024 is set to see AI playing an increasingly prominent role in all kinds of tech devices and services, and Google is getting the ball rolling by enhancing Google Assistant with Google Bard features, having launched its AI chatbot last year. 

During its Made by Google event in October, Google announced that the new Assistant by Bard would blend elements of both tools to create a generative AI search powerhouse. Its Google Assistant search tool has been integrated across the company’s products since its launch in 2016.  

Google’s developments in AI are transforming the way users experience and interact with its repertoire of apps and services, with AI tools available in Gmail, YouTube, and Google Docs, among others. The merging of Google Bard and Google Assistant features marks the next big step in the company’s plan to integrate AI across all its products and services. 

While Assistant with Bard doesn’t have a confirmed release date just yet, images and video shared by 9to5Google give us an idea of how it will look and function. 

9to5Google suggests that Assistant with Bard will replace Google Assistant altogether across Google and Android devices. If this is true, it’s likely that you’ll access the new AI the same way as you would access Google Assistant; either by commanding “Hey Google”, or long-pressing the power button. 

Looking at the images, the Discover page in the Google search app appears to have received a Bard integration in the form of a slider toggle that enables you to easily switch between a standard Google search and the AI chatbot

Assistant with Bard first look

(Image credit: 9to5Google )

Other images show the pop-up that appears when Assistant by Bard is enabled, allowing you to ask questions by talking, typing, or sharing photos using the three options at the bottom of the screen. Google previewed this design during its October event, at which it launched the Google Pixel 8 and Pixel 8 Pro.  

Assistant with Bard first look

(Image credit: 9to5Google )

Assistant with Bard isn’t yet available to use, but going by the images shared by 9to5Google it appears that the rollout of Google’s next AI development is imminent.  

You might also like

TechRadar – All the latest technology news

Read More

Google’s Nearby Share tool appears to adopt Samsung’s similar utility name and we wonder what’s going on

Google has suddenly changed the name of its file-sharing tool from Nearby Share to Quick Share which is what Samsung calls its own tool.

It’s a random move that has people scratching their heads wondering what it could mean for Android in the future. This update appears to have been discovered by industry insider Kamila Wojiciechowska who displayed her findings on X (the platform formerly known as Twitter). Wojiciechowska revealed that she received a notification on her phone informing her of the change after installing Google Mobile Services version 23.50.13. 

In addition to the new name, Google altered the logo for the feature as well as the user interface. The logo will now consist of two arrows moving toward each other in a half-circle motion on a blue background. Regarding the UI, it will now display a Quick Settings tile for fast configuration, text explaining what the various options do, and an easier-to-use interface. There’s even a new ability, allowing people to restrict Quick Share visibility down to ten minutes.

Wojieciechowska states this update is not widely available nor is the Nearby Share change common among the people who do receive the patch. This may be something only a handful will receive. She admits to being confused as to why Google is doing this, although it appears this could be the start of a new collaboration between the two companies according to found evidence.

Start of a new partnership

Android Authority in their report claims Wojieciechowska discovered proof of a “migration education flow” for Quick Share after digging through the Play Services app. This could suggest Google and Samsung are combining their file-sharing tools into one. Or at the very least, “making them interoperable”. 

If this is the case, two of the biggest Android brands coming together to unify their services could be a huge benefit for users. Currently separate and similarly behaving features might, if this is any evidence, coalesce into one that’ll work with both Galaxy and non-Galaxy smartphones alike. It's a quality-of-life upgrade that'll reduce software clutter.

Android Authority makes it clear, though, that there isn’t any concrete proof stating the two tools will merge. It’s just given the set of circumstances that seems to be the case. Plus, the whole thing wouldn’t make sense if it wasn’t the result of an upcoming collaboration. Think about it. Why would Google decide to give one of its mobile tools the same name as one of its competitor’s software? That might confuse users. 

There has to be something more to it so we reached out to both companies for more information. This story will be updated at a later time.

Until then, check out TechRadar's list of the best smartphone for 2023.

You might also like

TechRadar – All the latest technology news

Read More

Google’s AI-powered NotebookLM is now available to help organize your life

Google’s AI-powered writing assistant, NotebookLM, is leaving its experimental phase behind and launching as an official service with multiple performance upgrades.

First revealed during I/O 2023 as Project Tailwind, the tool’s primary purpose is to help organize your messy notes by creating a summary detailing what you wrote down. It will even highlight important topics plus come up with several questions to help people gain a better understanding of the material. The big update coinciding with the release is that NotebookLM will now run on Gemini Pro, which the tech giant states is their “best [AI] model” for handling a wide variety of tasks. It claims the AI model will enhance the service’s reasoning skills as well as improve its ability to understand the documents it scans.

What’s more, Google took the feedback from NotebookLM’s five-month testing period and has added 15 new features aiming to improve the user experience. 

Highlight features

The company highlights five specific features in its announcement with the first one being a “new noteboard space”. In this area, you’ll be able to take quotes from the AI chat or excerpts from your notes and pin them at the top for easier viewing. Next, citations in a response will take you directly to the source, “letting you see [a] quote in its original context.”

Highlighting text in said source will now suggest two separate actions. You can have the AI instantly “summarize the text” into a separate note or ask it to define words or phrases, which can be helpful if the topic is full of tough concepts. Down at the bottom, users will see a wider array of follow-up actions from suggestions on how to improve your prose to related ideas that you can add to your writing. NotebookLM will also recommend specific formats for your content that’ll shape it into an email layout or the outline of a script among other things.

NotebookLM sample

(Image credit: Future)

The full list can be found on Google's Help website. Other notable features include an increased word count for sources (they can now be 200,000 words total), the ability to share notebooks with others much like Google Docs, and support for PDFs.

Coming soon

There are more updates on the way. Starting on the week of December 11, NotebookLM will gain an additional seven features. They include a Critique function where you can ask the AI for constructive feedback plus a way to combine all your notes into one big page.

NotebookLM is available in the United States only to users ages 18 and up on desktop and mobile devices. When you visit, you’ll see some examples to help you get started with the service. It’s worth mentioning that despite this being an official launch, Google still regards NotebookLM as “experimental” technology, so it won’t be perfect. No word on if there are plans for an international release although we did ask. This story will be updated at a later time. 

While we have you check out Techradar's roundup of the best AI writers for 2023.

You might also like

TechRadar – All the latest technology news

Read More

Google’s Instrument Playground offers a taste of an AI-generated musical future

Google has opened the gates to its latest experimental AI called Instrument Playground which allows people to generate 20-second music tracks with a single text prompt.

If that description sounds familiar to you, that’s because other companies have something similar like Meta with MusicGen. Google’s version adds two unique twists. First, it’s claimed to be capable of emulating over 100 instruments from around the world. This includes common ones like the piano to more obscure woodwinds like the dizi from China. 

Secondly, the company states you can “add an adjective” to the prompt to give it a certain mood. For example, putting in the word “Happy” will have Instrument Playground generate an upbeat track while “Merry” will create something more Christmassy. It’s even possible to implement sound effects by choosing one of three modes: Ambient, Beat, and Pitch. For the musically inclined, you can activate Advanced mode to launch a sequencer where you can pull together up to four different AI instruments into one song.

Live demo

The Instrument Playground is publicly available so we decided to take it for a spin.

Upon going to the website, you’ll be asked what you want to play. If you’re having a hard time deciding, there is a link below the prompt that opens a list of 65 different instruments. We said we wanted an upbeat electric guitar, and to our surprise, the AI added backup vocals to the riff – sort of. Most of the lyrics are incomprehensible gibberish although Chrome’s Live Caption apparently picked up the word “Satan” in there.

The generated song plays once (although you can replay it at any time by clicking the Help icon). Afterward, you can use the on-screen keyboard to work on the track. It’s not very expansive as users will only be given access to 12 keys centered around the C Major and C Minor scales. What you see on the page is directly tied to the numbers on a computer keyboard so you can use those instead of having to slowly click each one with a mouse.

Instrumental Playground example

(Image credit: Future)

You can use the three modes mentioned earlier to manipulate the file. Ambient lets you alter the track as a whole, Beat highlights what the AI considers to be the “most interesting peaks”, and Pitch can alter the length of a select portion. Users can even shift the octave higher or lower. Be aware the editing tools are pretty rudimentary. This isn’t GarageBand.

Upon finishing, you can record an audio snippet which you can then download as a .wav file to your computer. 

In the works

If you’re interested in trying out Instrument Playground, keep in mind this is an experimental technology that is far from perfect. We’re not musicians, but even we could tell there were several errors in the generated music. Our drum sample had a piano playing in the back and the xylophone sounded like someone hitting a bunch of PVC pipes. 

We reached out to Google with several questions like when will the AI support 100 instruments (If you remember, it’s only at 65 at the time of this writing) and what the company intends to do with it. Right now, Instrument Playground feels like little more than a digital toy, only capable of creating simple beats. It'd be great to see it do more. This story will be updated at a later time.

While we have you, be sure to check out TechRadar's list of the best free music-making software in 2023

You might also like

TechRadar – All the latest technology news

Read More

Google’s AI plans hit a snag as it reportedly delays next-gen ChatGPT rival

Development on Google’s Gemini AI is apparently going through a rough patch as the LLM (large language model) has reportedly been delayed to next year.

This comes from tech news site The Information whose sources claim the project will not see a November launch as originally planned. Now it may not arrive until sometime in the first quarter of 2024, barring another delay. The report doesn’t explain exactly why the AI is being pushed back. Google CEO Sundar Pichai did lightly confirm the decision by stating the company is “focused on getting Gemini 1.0 out as soon as possible [making] sure it’s competitive [and] state of the art”. That said, The Information does suggest this situation is due to ChatGPT's strength as a rival.

Since its launch, ChatGPT has skyrocketed in popularity, effectively becoming a leading force in 2023’s generative AI wave. Besides being a content generator for the everyday user, corporations are using it for fast summarization of lengthy reports and even building new apps to handle internal processes and projections. It’s been so successful that OpenAI has had to pause sign-ups for ChatGPT Plus as servers have hit full capacity.

Plan of attack

So what is Google’s plan moving forward? According to The Information, the Gemini team wants to ensure “the primary model is as good as or better than” GPT-4, OpenAI’s latest model. That is a tall order. GPT-4 is multimodal meaning it can accept video, speech, and text to launch a query and generate new content. What’s more, it boasts overall better performance when compared to the older GPT-3.5 model, now capable of performing more than one task at a time.

For Gemini, Google has several use cases in mind. The tech giant plans on using the AI to power new YouTube creator tools, upgrade Bard, plus improve Google Assistant. So far, it has managed to create mini versions of Gemini “to handle different tasks”, but right now, the primary focus is getting the main model up and running. 

It also plans to court advertisers with their AI as advertising is “Google’s main moneymaker.” Company executives have reportedly talked about using Gemini to generate ad campaigns, including text and images. Videos could come later, too.

Bard upgrade

Google is far from out of the game, and while the company is putting a lot of work into Gemini, it's still building out and updating Bard

First, if you’re stuck on your math homework, Bard will now provide step-by-step instructions on how to solve the problem, similar to Google Search. All you have to do is ask the AI or upload a picture of the question. Additionally, the platform can create charts for you by using the data you enter into the text prompts. Or you can ask it to make a smiley face like we did.

Google Bard's new chart plot feature

(Image credit: Future)

If you want to know more about this technology, we recommend learning about the five ways that ChatGPT is better than Google Bard (and three ways it isn't).

Follow TechRadar on TikTok for news, reviews, unboxings, and hot Black Friday deals!

You might also like

TechRadar – All the latest technology news

Read More