Researchers at UCSD have successfully implemented DFlash, a block-diffusion speculative decoding method, on Google TPUs to bypass the sequential bottlenecks of traditional autoregressive drafting. By "painting" entire blocks of candidate tokens in a single forward pass rather than predicting them one-by-one, the system achieved average speedups of 3.13x, with peak performance nearly doubling that of existing methods like EAGLE-3. This open-source integration into the vLLM ecosystem optimizes TPU hardware by leveraging "free" parallel verification and high-quality draft predictions for complex reasoning tasks.

Karrot is a hyperlocal, community-driven peer-to-peer marketplace app that enables users to buy, sell, and trade items with other verified users. Since launching in South Korea in 2015, the platform has expanded into global markets, amassing over 43 million registered users.

After launching in North America, engineers at Karrot observed that 30% of users in the region use a non-English device language, such as Spanish. To make the app more accessible, the team wanted to bring seamless translation functionality to Karrot quickly and at scale. The developers determined that the most efficient way to implement quality translations would be through integrating an AI service directly into the app, so they selected the Firebase AI Logic and its Android SDK to access Gemini Flash Lite, which led to higher purchasing conversion among non-English users.

Integrating Gemini Firebase AI Logic

The team initially tested two on-device options: the ML Kit Translation SDK and Gemini Nano. But the team found challenges with each: ML Kit Translation didn’t meet the team’s quality expectations, and Gemini Nano, if it isn’t already on the device, required the user to download the model data.

The team then tested Firebase AI Logic. By calling the Gemini API directly from the app, Firebase AI Logic delivered accuracy at speeds that mirrored a natural conversational cadence.

Integrating Firebase AI Logic into the app was a “remarkably straightforward experience,” according to TaeGyu An, an Android Software Engineer on Karrot’s Mobile Platform team. TaeGyu and the team used the platform’s documentation and code samples to build a proof of concept in under three hours.

This allowed the team to spend more time refining prompts and finding optimal configuration values. “Even without extensive experience writing prompts, the official documentation's guides and tips made it easy to quickly identify the right direction for improving translation quality,” said WonJoong Lee, an Android Software Engineer on Karrot’s North America Product Team.

This low barrier to entry and rapid turnaround time enabled engineers to keep development costs low and go from proof of concept to production code in just two weeks—all without setting up a dedicated backend. That also freed up time to focus on UX and policy design, such as opt-in behavior and the conditions for the translation banner.

Driving sales with enhanced AI features

Since implementing translation using Gemini and Firebase AI Logic, the Karrot team observed higher purchasing conversion among non-English users, indicating that the translation feature is helping drive sales.

Of users who used a non-English device language, one in three of them who were shown the translation banner actively used the feature. The team has also observed that buyers offered translation functionality were 2.4X more likely to start a chat with a seller than those who weren’t.

The flexibility and simplicity of deploying Firebase AI Logic has led the team to explore other features to simplify the workstreams of its engineers. “It’s rewarding to build features that scale across diverse Android devices while helping neighbors connect and interact within their local communities,” concluded TaeGyu.

Going forward, the team plans to implement Server Prompt Templates to adjust prompts after release without shipping a new version of the app. This, combined with Remote Config, should help the team iterate faster and reduce operational overhead.

Get started

Learn how to build Gemini-enabled features like AI translations and in-app personalization and more with Firebase AI Logic to deliver better experiences to your users, faster.

Here are Google’s latest AI updates from April 2026

Previously available in beta, client-side encryption (CSE) customers can now conduct bulk migrations of sensitive files from both cloud and on-premises data sources. This process ensures confidential content is wrapped with customer-managed keys before it's imported into Google Workspace. Using this tooling, CSE customers can decommission their legacy third-party storage while ensuring the CSE model is in place throughout the document lifecycle. The API is highly configurable, we share a generalized sample code to make deployment simple, but customers can further customize it depending on their needs.

Getting started

Rollout pace

Availability

Resources

During National Small Business Week, jumpstart your growth with tools like Gemini, Workspace and Ads — plus, resources and exclusive offers from Google.

Event-Driven Webhooks are a push-based notification system that eliminates the need for inefficient polling.

Together with the National Archives and others, we’re launching an interactive journey through America’s history.

The Google for Startups Accelerator offers mentorship and technical support to companies using AI to disrupt the energy sector.

Starting June 1, 2026, Google Ads and related measurement APIs will transition to a 37-month data retention policy for granular performance statistics (daily, hourly, and weekly). High-level data (monthly, quarterly, and yearly) will continue to be available for 11 years.

API	Impact	Next Steps
Google Ads API and Google Ads scripts	Starting June 1, 2026, queries that request granular segments (such as segments.date, segments.week) for ranges older than 37 months from the current date will return a DateRangeError.INVALID_DATE error. Future API versions will return DateRangeError.REQUESTED_DATE_GRANULARITY_NOT_SUPPORTED error. To retrieve data older than 37 months, you must update your queries to use segments.month, segments.quarter, or segments.year. Unsegmented queries for historical data must align perfectly with calendar month boundaries (1st to last day of month) to succeed.	Please review your applications and make updates. If you require granular historical data beyond 37 months, we recommend exporting it prior to the June 1, 2026 deadline. If you have any questions, you can contact us on our Google Ads API support channel or Google Ads scripts support channel.
Google Analytics Data API	The Google Analytics Data API will truncate affected metrics to the latest 36-month window if the dimension "date" is also used in the report. Affected metrics include Advertiser Ad Cost, Clicks, and Impressions. Only reports with all of affected metrics, 37-months or older, and including the dimension "date" are impacted. Date-equivalent dimensions like "Day of week" and "day" are also impacted.	Review your data stitching logic to account for this truncation.
DV360 API and CM360 API	These APIs currently maintain a 24-month retention period, which remains unchanged.	No impact.
BigQuery Data Transfer Service	Starting June 1, 2026, the BigQuery Data Transfer Service for Google Ads and BigQuery Data Transfer Service for Search Ads 360 connectors will stop populating data for backfill runs with dates older than 37 months from the current date. Data transferred and stored in BigQuery will remain in the tables with no impact. Starting June 1, 2026, BigQuery Data Transfer Service for Google Analytics 4 connector will stop populating data for backfill runs with dates older than 37 months from the current date. Data transferred and stored in BigQuery will remain in the tables, but if a transfer is manually triggered for a report date 37-months or older, the data of the date in BigQuery table will be overwritten by empty value.	If you want to keep historical data beyond 37 months in BigQuery, we recommend starting backfilling runs early so that they can complete before June 1, 2026.

If you have any questions or want to discuss this post, please reach out to us on our “Google Advertising and Measurement Community” Discord server.

Nadine Wang, Advertising and Measurement APIs Team

Move from conversation to creation with file generation in Gemini

Students can now create personal class notebooks with NotebookLM in Google Classroom

Workspace audit logs: New functionality and expanded event fields in the Admin console

New ways to customize AI-generated meeting notes

New: Agent tools and security updates for Google Workspace developers

Google Takeout Transfer now supports Google Photos for Education users