Google Gemini can remember things now

Google has taken a major leap forward in personalized AI technology with its Gemini chatbot. The latest update allows Gemini to “remember” key details about its users, tailoring responses to their unique preferences and interests. Here’s what you need to know about this exciting development.

What Is Google Gemini’s New Memory Feature?

Imagine chatting with an AI that feels like it truly knows you, not in a creepy way but like a thoughtful friend who remembers your favorite coffee order or your passion for hiking. That’s the promise of Google Gemini’s new memory capability.

This feature, rolling out to subscribers of Gemini Advanced, enables the AI chatbot to remember your personal details and interests. Whether you’re a vegetarian looking for recipes or a fashion influencer needing professional tips, Gemini can now customize its responses to suit your needs.

How Does It Work?

  • Share Your Preferences: During a chat, you can share details about your work, hobbies, or aspirations. For example, you might tell Gemini, “I love gardening,” or “I’m studying to become a software developer.”
  • Use the ‘Saved Info’ Page: You can manually add or update details like dietary preferences, career interests, or favorite activities.
  • Edit or Delete Anytime: Not sure if you want Gemini to remember something? You can view, edit, or delete the saved data from the “Saved Info” page.
  • Transparency in Action: Gemini will notify you when it uses your saved information, keeping you in control.

Why Personalization Matters in AI

Personalization has always been the holy grail for tech companies. It’s what makes an AI feel less like a machine and more like a helpful assistant. By remembering your preferences, Gemini aims to:

  • Save time by delivering relevant suggestions.
  • Offer a more intuitive and engaging user experience.
  • Build trust by being transparent about how it uses your information.

Real-Life Scenarios

  • For Foodies: Gemini could suggest vegetarian recipes or gluten-free options based on your dietary choices.
  • For Professionals: Accountants might get tailored advice on tax-saving strategies, while writers could receive grammar tips.
  • For Hobbyists: Whether you’re into photography or marathon training, Gemini can guide you with personalized tips.

How Does Gemini Compare to ChatGPT?

Google isn’t the first to introduce memory features in an AI chatbot. OpenAI’s ChatGPT launched a similar feature in April 2024, available to ChatGPT Plus subscribers. Like Gemini, ChatGPT can recall details from past conversations, such as your favorite vacation spot or your pet’s name.

However, Gemini stands out for its integration within the Google ecosystem, potentially leveraging a wealth of contextual data (with user consent) to deliver even more precise recommendations.

Key Differences Between

Feature Google Gemini OpenAI ChatGPT
Memory Availability Gemini Advanced subscribers only ChatGPT Plus subscribers only
Language Support English (for now) Multiple languages
Transparency Notifies when using saved data Similar notification system
Integration Built into Google services Standalone

Is Privacy a Concern?

Whenever AI systems collect and store user data, privacy becomes a hot topic. Google has emphasized user control, allowing you to manage or delete saved data anytime. Transparency is also a focus, with notifications ensuring you’re aware of when your data is being used.

Still, it’s crucial for users to stay vigilant. Always review privacy policies and understand how your data will be handled.

How To Get Access

Gemini’s memory feature is currently exclusive to Gemini Advanced subscribers through the Google One AI Premium subscription. It’s available in English only, though additional languages may be added in the future.

Subscription Benefits

  • Access to advanced AI features.
  • Personalized responses tailored to your preferences.
  • Integration with Google services for seamless use.

What’s Next for Google Gemini?

The addition of memory is just the beginning. As AI continues to evolve, we can expect Gemini to expand its capabilities, offering even more intuitive and human-like interactions.
For now, this feature positions Gemini as a serious competitor in the AI chatbot space, giving users a personalized, user-friendly experience that feels ahead of the curve.

React 19: New features and Updates

React 19 is officially released on April 25, 2024, marking a significant milestone. This release brings a variety of new features and improvements designed to enhance developer experience and application performance. Many features that were experimental in React 18 are now considered stable in React 19, offering a more robust and optimized framework.

React 19 new features

React 19 release brings a host of groundbreaking features that significantly enhance the development experience and application performance. Let’s explore four key innovations:

1. Experimental React compiler: 

This new compiler transforms React code into optimized JavaScript, offering substantial performance gains. It autonomously manages component rendering for UI state changes, eliminating the need for manual state and UI updates.

The compiler also optimizes components as needed, reducing dependency on hooks like useMemo, memo APIs and useCallback. With React compiler, developers can handle errors encountered due to excessive re-rendering through performance optimizations.

Key benefits:

  • Improved error reporting in react-dom
  • Seamless rendering of async scripts anywhere in a component tree
  • Native rendering of document metadata tags in components
  • Better APIs loading and preloading browser resources to optimize resource loading

2. Server actions: 

The server actions streamline data mutations and state updates in React applications by bridging the gap between client and server seamlessly. This new enhancement of React 19 enables developers to automatically handle pending states, error management, optimistic updates, and sequential requests.

Key benefits:

  • Client components to call asynchronous functions
  • Provide built-in error handling
  • Enable immediate user feedback during data submission
  • Simplify form handling through new props like action and formAction

3. React Server Components (RSCs): 

RSCs allow server-side rendering of UI components, separate from client-side applications or traditional SSR. This approach significantly improves initial page load times, especially for complex UIs. RSCs can be executed at build time or for each request, offering flexibility.

They reduce client bundle sizes and integrate seamlessly with Client Components, creating a balanced architecture that combines server-side efficiency with client-side interactivity.

Key benefits:

  • Faster page load time and smoother interactions
  • Runtime flexibility at build time, and for each request
  • Improved app performance

4. New Hook: useFormStatus, useActionState, useOptimistic  

The newest React 19 version comes with three prominent hooks as significant additions in the library. These React’s built-in hooks are powered by React programming patterns. In React 19, these new hooks are going to transform the way React developers handle forms.

  • useActionState 

UseActionState is designed to simplify state management within Actions. It automatically tracks pending states during an action and returns both the final result and a pending status indicator. useActionState also supports action composition, managing execution order and state updates for chained actions.

Key benefits:

  • Automatically manages pending state 
  • Eliminates manual handling
  • useFormStatus

React 19 introduces useFormStatus, a hook that simplifies access to parent form information within child components. This is particularly valuable in design systems where components need to interact with form state without excessive prop drilling.

Key benefits:

  • Provides direct access to parent form status for child components, similar to Context providers
  • Minimizes prop drilling, leading to cleaner component hierarchies
  • Addresses a common design system scenario, reducing boilerplate code for form state interactions
  • useOptimistic

The useOptimistic hook in React 19 facilitates optimistic UI updates during asynchronous data mutations. This pattern allows immediate display of expected outcomes while data updates are in progress.

Key benefits:

  • Enables instant optimistic state rendering, providing immediate user feedback
  • Automatically manages state, reverting to original values if updates fail
  • Enhances user experience by increasing perceived responsiveness and interactivity

5.  New API: use

React 19 release introduces an experimental API called use, designed to read resource values (like Promises or context) directly within render functions. This leads to more concise code for asynchronous data fetching and state management.

Key benefits:

  • Allows direct access to resource values in render functions, reducing the need for separate state variables or lifecycle methods
  • Primarily targets data fetching scenarios, simplifying Promise handling
  • Currently available in React’s Canary and experimental channels, with potential for future changes

New features in React 19: Notable improvements for developers at a glance

1. Ref as a Prop

React 19 introduces a notable enhancement for refs: functional components can now directly accept refs as props. This advancement often eliminates the need for the forwardRef higher-order component.

Key benefits:

  • Enhanced functional components: Refs can now be passed using standard prop syntax, increasing versatility when interacting with DOM elements.
  • Streamlined code: Removing the forwardRef requirement leads to cleaner, more concise component code and improved readability.
  • Maintains compatibility: forwardRef is still supported for scenarios involving class components or complex functional component hierarchies, ensuring existing code remains functional.

2. Improved hydration error reporting

 React 19 significantly enhances error reporting, particularly for hydration errors. This improvement aims to make debugging more straightforward and efficient.

Key benefits:

  • Clearer error messages: Instead of displaying multiple confusing errors, React now provides a single, comprehensive error message.
  • Detailed difference reporting: The new error format clearly highlights discrepancies between server-rendered and client-side content.
  • Enhanced debugging experience: Developers can more easily identify and resolve hydration-related issues, leading to smoother development processes.

3. Support for preloading resources

ditReact 19 enhances performance optimization with a new feature: built-in support for resource preloading. This capability allows developers to proactively load fonts, scripts, and stylesheets in the background while users interact with the initial page content.

Key benefits:

  • Enhanced Page Load Speed: Preloading enables browsers to fetch resources before they’re critically needed for rendering. This leads to a more fluid user experience and faster perceived load times.
  • Optimized User Experience: By reducing the amount of content downloaded after the initial page load, preloading creates a smoother experience for users, particularly beneficial for those on slower connections.
  • Declarative Resource Management: React 19 provides access to modern browser APIs such as ‘preload’, ‘prefetch’, and ‘preconnect’. This allows developers to declaratively control resource preloading within their components, offering a more intuitive and React-centric approach to performance optimization.

This feature represents a significant step forward in React’s built-in performance tooling, giving developers more control over resource loading strategies directly within their component code.

4.  Async Scripts for better asset loading 

React 19 introduces enhanced support for asynchronous scripts, significantly improving how these scripts are managed within component hierarchies. This update streamlines the handling of scripts that load asynchronously and in potentially variable order throughout the component tree.

Key benefits:

  • Versatile Script Positioning: Developers can now place async scripts anywhere in the component tree without concerns about duplication or misplacement. React automatically ensures that each script is loaded and executed only once, even when rendered by multiple components.
  • Simplified Code Structure: The need for complex custom logic to manage script loading order or prevent duplication is eliminated. React now handles these aspects internally, resulting in cleaner, more concise code.
  • Enhanced Code Organization: By allowing async scripts to be positioned near the components that depend on them, this feature promotes more intuitive and maintainable code structure. This proximity improves code readability and eases future modifications.

This improvement in async script handling represents a significant step towards more efficient and developer-friendly management of external resources in React applications, reducing complexity and potential errors associated with script loading.

5. Style support for document metadata 

React 19 brings a notable advancement for i.e. seamless support for document metadata tags under components. With this update, React developers now add relevant metadata tags in rendering documents minimizing the hassle of manual tags insertion. In React 19, developers get meta tags like <title> <link> and <meta> under the <head> section of the document when React renders this component. The metadata tags bridge all the disruption gaps across client-only apps, streaming SSR, and Server Components. 

Key benefits:

  • Improved SEO optimization: Particularly useful for SEO purposes, the document metadata support in React 19 enables modifications in meta tag strings, which are meant to increase website visibility in search engines.
  • Enhanced social media sharing: Make it simple to quickly edit and tweak Open Graph and Twitter Card elements, leading to much better and more accurate previews when content is shared on social media platforms.
  • Better accessibility: Support for document metadata may also help in the management of ARIA attributes and other types of metadata that contribute to better accessibility of the developed React apps.

What is OpenAI o1? And how does it compare to GPT-4o?

OpenAI announced the release of its new series of AI models—OpenAI o1, with significantly advanced reasoning capabilities. According to OpenAI, what sets the o1 apart from the GPT-4o family is that they’re designed to spend more time thinking before they respond. One of the caveats with older and current OpenAI models (e.g. GPT-4o and 4o-mini) is their limited reasoning and contextual awareness capabilities—which lag behind advanced models like Anthropic’s Claude 3.5 Sonnet. OpenAI o1 is designed to help users complete complex tasks and solve harder problems than previous models in science, coding, and math.

This blog explores Open o1’s features, test results, pricing, and comparisons with existing benchmarks, GPT-4o and Claude 3.5 Sonnet (you can compare currently leading models here)

1. Overview of OpenAI o1

OpenAI o1 is a model family designed specifically for advanced reasoning and problem-solving. According to Open AI, the models can perform similarly to PhD students on challenging benchmark tasks in physics, chemistry, and biology. The test results don’t suggest otherwise. Key highlights of the OpenAI o1 models include.

1.1 Performance Metrics

OpenAI o1 ranks in the 89th percentile on competitive programming questions and has shown remarkable results in standardized tests, outperforming human PhD-level accuracy in physics, biology, and chemistry benchmarks. Besides, the model has a 128K context and an October 2023 knowledge cutoff.

1.2 o1 Model Family

The series includes the o1 preview model with a broader world knowledge and reasoning, and a smaller variant, o1-mini, which is faster and more cost-effective, especially for coding tasks. The o1-mini is approximately 80% cheaper than the o1-preview while maintaining competitive performance in coding evaluations.

1.3 Availability of o1 models

The o1 preview models are currently available in ChatGPT Plus (including access for Team, and Enterprise users), as well as via API for developers on tier 5 of API usage. In ChatGPT, it has a strict message limit of only 30 messages per week for the o1 preview and 50 messages per week for the o1 mini, after which you are required to switch to GPT-4o models.

1.4 Pricing for o1 models compared with GPT-4o

OpenAI has structured its pricing to cater to different user needs, with o1-mini being the most economical option. Here’s a breakdown of the pricing for the OpenAI o1 models:

Model             Input Tokens                  Output Tokens
OpenAI o1            $15.00 / 1M $60.00 / 1M
OpenAI o1-mini            $3.00 / 1M $12.00 / 1M
GPT-4o (08-06)            $2.5 / 1M $10.00 / 1M
GPT-4o mini            $0.150 / 1M $0.600 / 1M
Claude 3.5 Sonnet            $3.00 / 1M $15 / 1M

2. Comparison of OpenAI o1 vs GPT 4o

In rigorous testing, OpenAI o1 has demonstrated superior reasoning skills compared to its predecessors. For example, in a qualifying exam for the International Mathematics Olympiad, the o1 model scored 83%, while GPT-4o only managed 13%. Additionally, the o1 model scored significantly higher on jailbreaking tests, indicating a stronger adherence to safety protocols.

Performance Comparison

The below charts (courtesy: OpenAI) provide some interesting details about OpenAI o1’s technical performance across different metrics:

3. Comparison of OpenAI o1 vs Claude 3.5 Sonnet

Here are some quick points highlighting the differences comparing OpenAI o1 with GPT-40 and Claude 3.5 Sonnet:

  • Reasoning Ability: OpenAI o1 outperforms GPT-4o in complex reasoning tasks, as evidenced by its superior scores in competitive programming and math challenges. But the context is still lower than Claude’s most premium plan aka Claude for Enterprise, which has a 500K context window.
  • Safety and Compliance: OpenAI o1 has shown improved performance in safety tests, indicating better adherence to safety protocols compared to GPT-4o and Claude 3.5 Sonnet.

Claude AI also launched its own github integration to ground the responses in your personal data, which is especially helpful for code generation use cases.

4. Conclusion

The introduction of OpenAI o1 marks a significant milestone in AI development, particularly in enhancing reasoning capabilities for complex problem-solving. OpenAI mentioned that they expect to add browsing, file, and image uploading, and other features to make them more useful to everyone. It’ll be interesting to follow along with these developments.  At the same time, it is important to compare models and pick the one which works best for your use case, and the most expensive model isn’t always the best. The best models currently are GPT-4o, Claude 3.5 Sonnet, Llama 3.1 and you can test multiple models and make a decision that works for you.