Google Bard Overhaul: App Coding Revolution via Image Screenshots

11 Mar 2024 · 11 min

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

AI Today Podcast: Episode Summary

Podcast Details

  • Title: AI Today
  • Description: This podcast explores the latest advancements and ethical considerations in artificial intelligence, providing insights into its impact on different industries and society.

Episode Overview

  • Episode Title: Google Bard Overhaul: App Coding Revolution via Image Screenshots
  • Description: This episode discusses a significant update to Google Bard, which now allows developers to code applications directly from image screenshots, marking a major shift in app development processes.

---

Key Takeaways

Major Updates to Google Bard

  1. Image Input Feature:
  2. Users can now upload images and receive descriptive responses.
  3. Potential applications include assisting individuals with visual impairments and coding from visual prompts.
  4. Concerns raised about misuse, such as bypassing CAPTCHAs.
  1. Expanded Language Support:
  2. Google Bard now supports over 40 languages, enhancing its accessibility globally.
  3. Still not available in Canada, which poses challenges for users.
  1. Conversation History and Organization:
  2. Users can now access previous conversations and pin important threads for easier navigation.
  3. Suggestion for a future folder system to better organize prompts.
  1. Text to Speech Functionality:
  2. Users can listen to Bard's responses in various languages, improving accessibility.
  1. Sharing Conversations:
  2. Users can share their Bard conversations, allowing others to view useful interactions.
  1. Adjustable Response Styles:
  2. Users can modify Bard's responses based on desired tone (e.g., casual, professional).
  1. Enhanced Code Exporting:
  2. Ability to export Python code to multiple platforms including Replit and Google Colab.
  1. Modifiable Responses:
  2. Users can edit Bard’s responses straight away, allowing for personalized outputs.

---

Innovative Use Cases

  • App Development from Images:
  • Users can upload screenshots of applications and receive corresponding code.
  • Example: A user recreated an iPhone app from a screenshot.
  • Medical Diagnosis:
  • Google Bard was used to analyze a brain CT scan, showcasing its potential in healthcare applications.
  • Recipe Generation:
  • Users uploaded food images, and Bard provided relevant recipes, demonstrating its utility in culinary tasks.

---

Industry Impact

  • The advancements in Google Bard position it as a strong competitor to OpenAI's GPT-4, particularly in integrating visual prompts into coding and application development.
  • The episode highlights the rapid pace of improvement in AI tools and raises questions about how established models like ChatGPT will respond to these innovations.

---

Conclusion The introduction of these features to Google Bard represents a significant leap in AI capabilities, especially in the realms of accessibility and app development. As Bard continues to evolve, it could reshape how developers and users interact with AI, creating new opportunities and challenges in the tech landscape.

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:00The wait is over. Dive into Audible's most anticipated collection, The Best of 2025. featuring top audiobooks, podcasts, and originals across all genres. Our editors have carefully curated this year's must-listens from brilliant hidden gems to the buzziest new releases. Every title in this collection has earned its spot. This is your go-to for the absolute best in 2025 audio entertainment. Whether you love thrillers, romance, or nonfiction, your next favorite listen awaits. Discover why there's more to imagine when you listen at audible.com slash best of the year. What can 160 years of experience teach you about the future?

0:44When it comes to protecting what matters, Pacific Life provides life insurance, retirement income, and employee benefits for people and businesses building a more confident tomorrow. Strategies rooted in strength and backed by experience. Ask a financial professional how Pacific Life can help you today. Pacific Life Insurance Company, Omaha, Nebraska, and in New York. Pacific Life and Annuity, Phoenix, Arizona. Today, Google Bard has released some absolutely massive new features that make it better than ChatGPT and GPT-4. I'm sorry to say, I've always been a long-term, you know, OpenAI GPT fan.

1:21I have their premium. But Google Bard really just did something very powerful that I think is giving not just OpenAI a run for their money, but I genuinely believe it makes it a more useful and more powerful tool. So today on the podcast, I'm going to dive into the new features they've just added and what this means for the industry overall. So the first thing that you should know that they added is the ability for you to input images and get responses back. Now, the thing about this and a bunch of these other features is some of these things are things that OpenAI has announced that they're going to do with GPT-4.

1:55Like when they announced GPT-4 months ago, they said, pretty soon you're going to be able to, you know, draw a website on a napkin and put it in. We're going to code it for you. Like they said this, but then they never actually shipped it. And if there's anything that I am not a fan of, it's announcing new features and capabilities and not shipping them. Like, what's the point? Just wait till you're going to ship it to tell us, because I don't want to know months ahead, what is possible, allegedly, what I could have allegedly, but don't. So Google beat them to the punch on this feature which like let's be honest this thing was months ago that they announced allegedly it could do so i really think this is quite impressive um so essentially they you know i saw a screenshot of someone they said what color or what's the color of this cat they uploaded a picture of a cat google bar said uh the cat in the image is white with blue eyes blah blah blah it gives a bunch of other information about this this is awesome for people with um visual impairments and disabilities to be able to see stuff.

2:53Of course, that's like one version, but people are saying that this is really impressive for a lot of different reasons. I mean, theoretically, you could screenshot a CAPTCHA. Oh man, I didn't even think about this implication. You could screenshot a CAPTCHA, put it in here and say, what are the numbers and letters in this picture? It could tell you, and you could just build a bot or have, you know, hackers or spammers or scammers or whoever completely kill CAPTCHAs with a tool like this. It does a lot more than just killing CAPTCHAs, But overall, I think that is, you know, one very interesting use case.

3:23There'll be many, many more. The next feature that they recently launched is expanded language. So you're going to be able to actually use Google Bard in over 40 language, including Arabic, Chinese, German, Hindi, Spanish. And I think this is going to be really impressive. You're going to be able to access Google Bard in more places now with their recent update, including Brazil and across Europe. still not in Canada which is a really big bummer because I'm on vacation in Canada for the next month or so and so I got to use a VPN every time I want to use this or other AI tools which my opinion is absolutely ridiculous but that's another topic for another time the other feature that's really interesting is you can now access previous conversations with Bard before it was just like you chat with it you come back and it's gone so this is good this is similar to chat GPT this is whatever this just should have been here from the beginning but happy that they have this now something that is interesting is that you can actually pin your most important threads to make them easier to access and to kind of organize it better I actually do think this is interesting this is kind of annoying uh with with chat gpt like sometimes I'm like trying to scroll down and figure out what thread it was the way it's auto labeled it's just it's not my favorite and so um being able to pin stuff I think that's a good first step honestly there just needs to be a folder system there where you can drag prompts you can make folders and organize it like crazy I think that would be an absolute game changer but I'm sure it'll get there eventually the other one that Google bar now has is text to speech capabilities so you can listen to a response in over 40 languages and I think this is really really cool because it really kind of expands your ability to explore the ideas if you get it to write a poem for you you can listen to what that poem actually sounds like when someone's reading it out loud.

5:05Again, good for people with different impairments to being able to actually just like listen to it being read back to you immediately right there. I'm curious to see if they'll do like an export audio feature. I think that might be really interesting would have a ton of implications, right? That would be something similar to what 11 labs is doing where they allow you to do text to audio and actually export the files to use them for whatever you want. The next feature that I think is really impressive is that you're able to, I mean, actually not really impressive. This is definitely hyperbole, sort of useful is that you can share conversations with friends.

5:41So there's before this been a lot of third party apps with things like, I think there's one called ShareGPT that got kind of popular and ChatGPT launched. But this is useful. Like you can save the conversation you've had, share with a friend so they can actually read through it. This is really useful for a lot of different things, you figure out how to do something cool, or you get a cool response, and you want to share that piece of information, this is a great way to do it. The next one is that you can easily adjust Bards responses. So you can actually change the tone and style of Bards responses to five different options.

6:10You have simple, long, short, professional, or casual. Personally, I love this because I have a lot of different prompts in ChatGPT where I'm like, you know, do this in a very professional tone, or do this in a casual tone, or make sure you're only using three sentences in this response you know like trying to get the exact thing and so it's kind of cool that like if you need the same type of response a whole bunch of times in a row you could just go and set the default option to one of those which is is pretty useful to be honest another one that is fairly interesting is the ability to export code to more places so google bar is going to allow you to export python code to replet in addition to google colab i think this feature is now live in over 40 languages which is pretty awesome another thing you can do is modifiable responses so google bard now allows you to modify its responses according to your requirements and needs right so um it gives you a response and you can go and edit that response right there i think that's going to be really powerful because it's going to help it's like essentially you're training it to know exactly what you want as a response so i think that's and give it some great data.

7:16Another thing you can do is you now have the ability, well okay so this is kind of another feature we already covered but essentially with the one where you can upload an image and it can tell what's inside the image, you can do more than that. Like the implications of being able to upload images and use images as prompts gets a lot bigger. So you can do things like the napkin, the classic napkin one that ChatGPT announced that they never actually launched where you can upload an image of a website that you drew on a napkin and it can make your website some someone actually literally did it and it's like kind of cool um they there was another there was a couple other really cool implications for this someone did one where they uploaded an image and they said give me the recipe for this image and it went and found a really good recipe now this is something that i think google is going to absolutely crush it at like if you really think about it google already had like a feature so so a lot of this tech they already had they just weren't integrating it all and stringing it together and so i think that's why microsoft and open ai are going to kind of struggle but they already had the ability with google images to take a picture of something and it would like search google images for that photo so you can imagine um there's a few different ways that you could go about getting a recipe because i've actually asked chad should be t for recipes before and it's notoriously bad like the like the I don't know, like the ratios of ingredients is just not correct.

8:39But I think this actually is going to be a lot better because Google can essentially, they could do two things. They could take your image and search it on Google and see like the most similar images to that. And then they could go and they could go and see if those are attached to like recipe websites, or they could use their image identifier to identify what the image, like what food is in the image and then they can go search for recipes for that specific thing and share them and the way that google ai works is it's more like they go and find existing things and just like re-say them existing actually useful outputs whereas open ai is kind of like making random stuff from scratch based off of whatever's in its database with the 2021 cutoff so i actually think it might be better it would have better results um making a recipe off of an image you gave it so that's kind of cool another one that something did someone did was they created an iphone app from a screenshot so they literally went and screenshotted the timer app on their iphone they put it on and said um i believe that their prompt was something like recreate the app you've seen the screenshot make sure it's fully functioning provide all of the necessary code for it to work write your code in swift um and then they uploaded the image and it literally was able to go and code the app now the app wasn't absolutely perfect um so but there's a couple things they had to go and fix in it but like pretty much it did the whole thing i watched the process of them like tweaking the code there's a just a handful of small little tweaks they did um and the thing was fully running so i it's really really impressive they ran in xcode and i think this is like an awesome implication obviously that's a small app but like the amount of time that that would save you could be pretty astronomical another really interesting use case of the uploading the images is that someone actually got Google Bard to diagnose a brain CT image.

10:25So it was a scan of the brain. They uploaded it and they asked Google Bard for a diagnosis of like a tumor in it pretty much, which was crazy. They can also, other people have also done the same one where they essentially converted a screenshot of an image to code. They got someone to, you know, convert a screenshot to Jetpack Compose, which was pretty impressive. And yeah, overall, I just think, you know, the image one is going to be really big because they're implementing it with they're integrating it with code. So I think that's going to be absolutely insane. But there's a lot of really other really other interesting use cases that are starting to pop up on Twitter that we're finding.

11:06So overall, I think this is actually a very massive update for Google. I got to give them a hats off. I got to say Bard is officially better than GPT for the time being, in my opinion, with these integrations, and especially with the image upload, that's the biggest one. Google Bard has the capability, or ChatGPT has the capability of doing that, GPT-4, so now it's just left to be said when they will actually launch that. And of course, GPT is definitely better for developers or anyone building anything because their API is already out and is way easier to integrate. So they do have a leg up on some things, but it's gonna be interesting to see how this plays out and how they kind of fight off Google Bard as it's making these rapid improvements and making their product like very very quickly the best product on the market

From the publisher

In this episode, we explore the major update to Google Bard, delving into how developers can now code applications directly from image screenshots, revolutionizing the app development process and democratizing access to programming.

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

More from AI Today

All 897 episodes
Google Bard Overhaul: App Coding Revolution via Image ScreenshotsAI Today · 11 min
Listen in VO