#AI - Page 4
410 Stories
Google’s AI is Losing It! Asks Users to Eat Rocks, Add Glue to Pizza & More
Arjun Sha
View quick summary
Google's new AI Overview experience in Search is rolling out to users in the US. People are complaining about the misinformation AI Overview is generating. We have collated some of the replies generated by Google's AI Overview. Essentially, Google has redefined the relationship of a search engine provider and taken the role of a publisher.
You can easily install the new Copilot app on Windows 11, 10, and older builds as well. I have extracted the new Copilot app from the 24H2 build and installed it on my PC running the 23H2 build. The new Copilot app is actually a web-based app and not a native Windows app.
Google and Microsoft are Already Testing Ads in AI Overview and Copilot
Arjun Sha
View quick summary
Google and Microsoft are now testing to display ads in AI Overview and Copilot. Google is initially rolling out ads in AI-generated results in the US. Microsoft is also adding ads to its generated responses on Copilot.
For the first time, a mega-corporation like Microsoft has released an open-source multimodal model. The new Phi-3 Vision model is trained on 4.2B parameters and brings a context length of up to 128K tokens. Microsoft has also released Phi-3 Small (7B) and Phi-3 Medium (14B) models.
In a major push to kickstart the new era of AI PCs, Microsoft has announced a new Recall AI feature on Copilot+ PCs. It actively takes screenshots of your PC every few seconds to create a local semantic index. The processing is done locally and the information is encrypted. However, would you be comfortable actively sharing your screen for the Recall feature? We find out.
Last week, OpenAI demoed the 'Sky' voice during the ChatGPT 4o launch. The voice seemed straight out of the movie 'Her', where Scarlett Johansson lent her voice to the AI character 'Samantha'. Now, it turns out Sam Altman, the CEO of OpenAI, was in contact with Scarlett Johansson to voice the new ChatGPT 4o system. However, Johansson had refused the offer, way back in September 2023. Now, Johansson is seeking clarity and transparency as to how the 'Sky' voice was developed. Meanwhile, OpenAI has paused the 'Sky' voice on ChatGPT.
Gemini 1.5 Flash is an Underrated Gem You Need to Try Right Now: Here’s How
Arjun Sha
View quick summary
At the I/O 2024, Google unveiled many AI models, but Gemini 1.5 Flash remained under the radar. It's a lightweight AI model that delivers remarkable speed and efficiency with support for multimodal reasoning and a large context window of 1 million tokens. It's also very cheap to run. You can try the model on Google AI Studio for free and without any waitlist.
If you don't like the new AI Overview on Google Search, you can turn it off by following our guide. We have added two extensions to entirely block AI-generated answers on Google Search. Apart from that, we have added three other ways to remove AI Overviews on desktop, Android, and iOS.
If you are wondering whether you should subscribe to ChatGPT Plus or keep using the free ChatGPT version, read our extensive comparison. We have done a thorough comparison of ChatGPT 4o and ChatGPT 4 models. In addition, we have laid out the differences between the free and the paid version of ChatGPT.
If you want to create a simple game or a web app, ChatGPT 4o can be of immense help. It created an arcade game in Python and I could run it without any errors on the first try. ChatGPT 4o further improved the game with a replay functionality and added a score system as well.
We have thoroughly compared ChatGPT 4o with Gemini 1.5 Pro on a variety of tasks including reasoning, code generation, multimodal tests, and more. In our tests, ChatGPT 4o performed much better than the Gemini 1.5 Pro model. To understand the capabilities of both models, go through our entire comparison.
Ilya Sutskever bids farewell to OpenAI and steps down from the role of chief scientist at the company. Ilya says he will be working on a deeply meaningful personal project. Many reports suggest that Ilya was responsible for OpenAI CEO Sam Altman's firing in November 2023.
Google has announced its upcoming open-source model Gemma 2 and it will be trained on 27 billion parameters. The model will be released in June and Google claims that it can outperform some of the AI models twice its size.
At the Google I/O 2024 event, Google announced a new feature called Gemini Gems that lets you create a custom AI chatbot. You can create virtual AI assistants for many roles and characters. The feature will be available to Gemini Advanced users.
At the Google I/O 2024 event, the search giant added a new model to the Gemini family. The Gemini 1.5 Flash model is developed for high-frequency tasks that require minimum latency. It still brings multimodal capabilities and a context window of 1 million tokens.
Google has expanded the context window of the Gemini 1.5 Pro model from 1 million to 2 million tokens. Developers can access the large context window in private preview. And Gemini 1.5 Pro also comes to consumers via Gemini Advanced with a context window of 1 million tokens.
OpenAI introduced its GPT-4o model recently and demoed many new cool experiences. However, some interesting details about the model's capabilities have not been highlighted. In this article, I take a look at all the interesting applications of ChatGPT 4o and list out the fascinating capabilities of the model.
You can access and use the GPT-4o model right away without requiring any subscription. The latest model is available on the web, ChatGPT's Android and iOS apps, and the ChatGPT macOS app as well. We have added a link to download the new ChatGPT desktop app for macOS.
ChatGPT’s New Feature Will Make It Easy to Analyse Files from Cloud Storage
Sagnik Das Gupta
View quick summary
An X user going by the username @legit_rumors has now spotted that ChatGPT is set to get a new Context Connector feature. With this feature, ChatGPT users will now be able to directly fetch their online Google Drive and Microsoft One Drive files and analyze that data using the AI chatbot.
A new report now confirms that OpenAI is going to release its ChatGPT search engine on Monday (May 13). The report also says that the OpenAI search engine will include citations and source real-time information from the web. The search engine has now been indexed by Bing and DuckDuckGo as well.
Anshuman Jain
In this feature, we explore whether we can mimic and get the same features as these new AI gadgets offer on our smartphones using the help of generative AI apps. We discuss all the things you can do, with which app and also how you can do it.
OpenAI is reportedly working on a new search engine, an alternative to Google Search. And the company may launch it in May, just before Google I/O. Currently, the subdomain doesn't throw a 404 error, but displays "Not found".
You can now chat with Gemini right from Chrome's address bar. You just have to type "@" and select "Chat with Gemini" to summon the AI chatbot. Enter your prompt and hit Enter. It will take you to Gemini's portal where you can find the answer. The feature is currently available to Chrome desktop users only.
Hellboy AI Character Design Controversy: Is Using AI Really That Bad?
Shashank Shakya
View quick summary
It's no secret that a new Hellboy movie with the title of Hellboy: The Crooked Man is being made. In a recent update, accusations of using AI for character design have been made on the director of this movie which he has confirmed to be false and to be a major quoting error made by the media house who published the report. However, my question is, is the use of AI actually that bad? Read about what I think in this article.
OpenAI Finally Relents; You Can Disable Model Training in ChatGPT While Retaining Chat History
Arjun Sha
View quick summary
You can now opt-out of model training on ChatGPT via Settings and also retain chat history. Earlier, OpenAI used to disable chat history if you refused to share your data for model training. You can find the new data controls under ChatGPT Settings.
As we are moving towards the AI era, things are developing at a breakneck pace. In all of this, we have to be mindful of our privacy and how to protect it. In this article, we look at the privacy policies of popular AI chatbots and how companies handle private conversations. We have also discussed how you can minimize your data footprint and opt out of model training.
OpenAI is now making the Memory feature available to all paid ChatGPT Plus users. It can remember your preferences and key details about yourself to make the chat experience more personal. The feature is turned on by default, but you can disable it from the Settings page.
Beebom Staff
Nvidia DLDSR is a feature that can be enabled via the Nvidia Control Panel. It is present under Manage 3D Settings > Global Settings > DSR - Factors > DL Scaling. When enabled, DLDSR upgrades the quality of your games using AI-based super sampling. Effectively, it can turn the rendering resolution of the game higher than what is supported by your monitor. The feature can even be enabled alongside DLSS for even more AI-enhanced goodness.
Microsoft is on a roll with smaller language models. It introduced a new family of Phi-3 models with the smallest, Phi-3 Mini available right away. The smallest model is trained on 3.8B parameters and beats larger models like Gemma 7B, Mistral 7B, and Llama 8B.
Do you want to run AI models locally on your Android phone? Well, with the MLC Chat app, you can now download LLMs like Gemma, Phi-2, Mistral, Llama 3, etc., and chat with them natively on Android devices. The token generation is slow though, especially on older Snapdragon-based Android devices.
Meta released its Llama 3 models recently so we have taken the liberty to compare the 70B model with the flagship GPT-4 model. Surprisingly, Llama 3 performs as good as the GPT-4 model despite being a smaller model. In advanced reasoning tests, it demonstrates intelligence and does better than GPT-4 in following user instructions.
Meta has finally unveiled its next family of Llama 3 models in two sizes: 8B and 70B parameters. The company has launched its meta.ai portal where you can chat with Llama 3 models. Besides that, you can find Meta AI experiences on WhatsApp, Facebook, Instagram, Messenger, and more. Currently, the context length is 8K tokens, but Meta will be releasing larger models with increased context length and multimodal capability in the near future.
Microsoft is working on a Circle to Copilot feature that is similar to Circle to Search by Google. The feature is available to try out in the Edge browser on iPhone and iPad. The feature is still in development and there is no confirmation when it will come to Android devices.
If you are looking for an easy solution that lets you chat with your documents locally, you can check out AnythingLLM. It has a slick GUI interface and works even on consumer-grade PCs. No need to have high-end GPUs. Apart from that, you can ingest a variety of file formats including TXT, PDF, CSV, audio files, spreadsheets, and more.
Upanishad Sharma
Nobody is safe from the clutches of AI as Logitech has now launched the new M750 AI Edition along with the Logitech Prompt Builder software. Combining ChatGPT integration with recipes, the builder can be assigned as a shortcut with certain Logitech mouse and comes up in a neat pop-up window. I had lots of fun with the app as it let me rewrite text, send custom e-mails, and even draft witty replies all with the power of ChatGPT! The app is free to use but does require the Logi Options App+.
Google has finally given access to Gemini 1.5 Pro API to all users. You can generate the API key from Google AI Studio and start calling the API for both text and image inputs. The API access is currently free as part of a public preview. We have also added some coding examples so go through our guide to understand the documentation.
You no longer have to wait to access the Gemini 1.5 Pro model. Google has made the model generally available to all users via Google AI Studio. You can access a context window of 1 million tokens without paying any fee. In addition to that, Gemini 1.5 Pro can now process audio files too, apart from videos and images.
Google at the Cloud Next 2024 event made its Imagen 2 model generally available to enterprise customers. One of its new features is that it can generate short video clips aka live images of up to four seconds. In addition, the generation clips show different camera angles and motion. Users can also inpaint and outpaint images on Vertex AI.
Spotify's new AI Playlist feature allows users to create personalized playlists by entering a prompt, which the AI will use to generate a playlist based on the user's listening history. The feature is currently available for Premium members in Australia and the UK, and will soon expand to other regions.
Soon You Might Have to Pay to Access Google’s AI Search Features
Sagnik Das Gupta
View quick summary
Google announced its SGE (Search Generative Experience) last year, and now, we are getting to see this entire thing become "premium". In other words, you will have to pay for the AI features integrated within Google Search. For those unaware, generative AI in Google Search allowed you to get a summary of your query, thereby providing all the information you need right at the top. Now, the new Financial Times report suggests that you will soon need to pay the premium $19.99/month fee of Google One AI to make use of it.
Using Google Gemini, you can summarize YouTube videos instantly and save up time. On Android, head over to the Gemini app -> profile icon -> Extensions -> toggle on YouTube extension -> then paste the YouTube video link in the Gemini text box -> hit send. On iOS, go to the Google app -> switch to Gemini -> follow same steps as Android. You can also visit Gemini on the web using the browser of your choice and summarize YouTube videos quickly.
YouTube CEO Says OpenAI Could’ve Violated Platform Rules For Sora Training
Arjun Sha
View quick summary
Speaking to Bloomberg, the YouTube CEO, Neal Mohan said that OpenAI would violate YouTube's Terms of Service if the company used YouTube videos to train its Sora model. Google also uses some corpus of YouTube videos to train its Gemini model, but signs licensing contracts with individual creators. Earlier, the CTO of OpenAI, Mira Murati, refused to clarify whether Sora was trained on YouTube videos.