Christopher Carr

@sdmat123 @OfficialLoganK @timurkhakhalev Won't happen *this* year.

@OfficialLoganK @timurkhakhalev Props for the honest answer about a difficult problem. Do you see Google getting to grips with capacity over the next year or two or is this the new normal?

We just shipped a bunch of stuff to make it easier to scale with the Gemini API: - Automatic tier upgrades - Tier 1 -> Tier 2 now happens much faster (30 days post payment -> 3 days) and with less spend ($250 -> $100) - New billing account caps on each tier to limit over spend

@OfficialLoganK @timurkhakhalev You didn't put in large enough orders, early enough, with TSMC. And then someone thought it wise to set a side a large number of TPUs for your frickin competition, Anthropic.

@timurkhakhalev 503's are the only tough situation here, the update from today should make getting less 429's much easier. For 503's we are making TPU's as fast as possible but there are certain times of day / week when we hit global capacity spikes which causes us to have to shed traffic.

@PheonixYin13 @OfficialLoganK I feckin can't stand "thought leader" bots.

Finally, built-in tools that actually play nicely with function calling and context circulation. The real winner here is long-running agents — they can now stay grounded in real-world data (especially Maps) without constantly losing context or building fragile glue code. This update makes Gemini feel much more production-ready for agentic workflows.

Lots of great Gemini API updates shipping today 🛠️ 1. Built-in tools (search, maps, file search) now work with function calling 2. We now do context circulation with built-in tools for better model performance 3. Grounding with Google Maps now works with Gemini 3!!

@HaseebMir91 @OfficialLoganK @coffee That's how billing is set up.

@OfficialLoganK @coffee but why do we need to create project? to get our API keys

The bottleneck has so quickly moved from code generation to code review that it is actually a bit jarring. None of the current systems / norms are setup for this world yet.

@sidvick @OfficialLoganK @coffee It's oddly easy to hit the 1 million TPM input rate limit -- especially if you're using OpenClaw.

@OfficialLoganK @coffee I have a paid subscription yet it’s really difficult to get the API keys to work without getting an API rate limit error!

@Vituswriter @BustBigPharma What are you calling "TV"? Lol, broadcast television? Do you watch streaming series and films on your phone?

@BustBigPharma Doesn't bother me one bit. Haven't watched TV in 14 years. I encourage everyone to do the same. Oh, and I don't accept ANY prescriptions unless I research the need and the drug and approve it myself. Drop dead #Pharma

🏆 “Most Annoying TV Ads” didn’t make the list of award categories, but the obvious winner is: BIG PHARMA Endless predatory ads w/crazy side effects, as Americans pay RECORD $$$ for prescription drugs CONGRESS MUST: 1. Ban Pharma Ads 2. Codify Trump’s 'Most Favored Nation' Plan

@BustBigPharma May cause explosive anal bleeding. Explosive anal bleeding may lead to death. If you experience explosive anal death stop taking Ansplozia and talk to your doctor.

@NormieNFT @OfficialLoganK Is your project in the AI Studios dashboard? The cap setting is in Dashboard -> Spend.

@OfficialLoganK Not vertex API. I am using the Gemini API (generative language API as you correctly pointed out) from google cloud console (console.cloud.gogle.com). Can I set up a hard cap on that Gemini API key's usage?

Spend caps in the Gemini API, available starting today!! This is another step forward of many to give developers more control and peace of mind when building with Gemini. Please go set a cap and send us any feedback as you use them!

@OfficialLoganK It sure doesn't seem like you have enough TPUs for this. That Anthropic deal (@ThomasOrTK @sundarpichai ) may well prove to be one of the stupidest things Google has ever done.

I sat down with Miriam and David from the Google Maps team to talk through and play around with the new Google Maps experience, it was awesome to see Gemini in action and learn from the Maps team!

Introducing our biggest upgrade to @googlemaps since the original launch, featuring Ask Gemini (with personalization), Immersive Navigation, and much more!! 🗺️

@jonatanpallesen "Rape conviction rate" Perhaps poorer people are disproportionately *convicted* of rape, relative to wealthier people with better lawyers. .. hmm, seems like there might have been some evidence of that, of late. Also, why did this neo-Nazi pop up in my feed?

Immigrant groups who rape the most are also the ones we pay the most to live in our countries.

@OspreyFunds Mmkay -- to what end?

Weekly XRP Update RIPPLE EXPANDS INTO JAPAN WITH MAJOR FINANCIAL PARTNERS Ripple is teaming up with Mizuho Bank, SMBC Nikko, and Securitize Japan to strengthen XRPL adoption through asset tokenization and payment infrastructure. The move aligns with Japan’s crypto-friendly reforms and deepens Ripple’s presence in Asia’s evolving blockchain ecosystem. RIPPLE TAPS AWS BEDROCK TO SUPERCHARGE XRPL OPERATIONS Ripple is working with Amazon Web Services to integrate Amazon Bedrock into the XRP Ledger, aiming to streamline system diagnostics and enhance operational efficiency. The AI-driven collaboration underscores XRPL’s push toward enterprise-grade performance and next-gen infrastructure innovation. FXRP SPOT TRADING LAUNCHES ON HYPERLIQUID VIA FLARE Flare has launched FXRP, enabling XRP spot trading on Hyperliquid’s onchain orderbook for the first time. The integration enhances XRP’s DeFi utility while preserving decentralized custody and cross-chain interoperability with the XRP Ledger.

@ivnardini @OfficialLoganK @matt_is_nice @cramforce Wouldn't it be better if there were a single Gemini API, and one could toggle on whatever additional enterprise/GCP features may be required?

@pdxweb @OfficialLoganK @matt_is_nice @cramforce Hi Christopher, the TLDR is that those APIs serve different user needs. But we've recently simplified migrating between them. Both the Gemini Developer API and the Vertex AI Gemini API are accessible through the unified Google Gen AI SDK. ai.google.dev/gemini-api/doc…

Announcing: APIKeyBench, the new baseline for AGI Send agent to get an API key for Gemini, WhatsApp, and Microsoft Teams. If API key works, eval passes.

@FinderGuys @OfficialLoganK @googlemaps What gibberish.

Wow, Google finally slapped some AI on their map, congrats on catching up in 2026 😂Our Apple Maps has been silently owning the game for years while you're out here begging "ask me anything". Quality? Data? Nah fam, drop that trash map and experience real navigation on iPhone. Apple Maps 1 – Google's AI wannabe 0, next please! 📷

@OfficialLoganK @googlemaps You're already horribly capacity constrained. Maybe add some TPUs (or claw some back from Anthropic) first. Can we get a subscription tier between Pro and Ultra while you're at it? Higher quotas, without the 30TB of storage.

@qladyi Probably better than Iran protected Ali Khamenei.

#KyraMachida ABD'nin köpekliğini yapan Arap liderleri Amerika kendini koruyamıyor sizi nasıl koruyacak??

@GoogleAIStudio Seems the cap ignores discounts (e.g., AI Pro/Ultra Gen AI & Cloud credits).

we’ve just launched project spend caps for the Gemini API in AI Studio. visit the new dashboard to set a dollar amount for your maximum spend to keep your costs predictable.

@BeamerSoftware @GoogleAIStudio That seems backwards

@GoogleAIStudio The more predictable the costs become, the less I can use AI.

@bbizzle0629 @GoogleAIStudio There is. For both Android and iOS

@GoogleAIStudio I saw something last year saying that there was a Google AI Studio app in the works. Is this legit drill?

@lumiq0 @GoogleAIStudio Effing "thought leader" bots.

@GoogleAIStudio The fact that AI companies now need spend cap features tells you everything about where API usage is heading. Developers aren’t experimenting anymore — they’re building at scale.

@GoogleAIStudio I would prefer to pre-fund a wallet. The cap being however much money I put in there. Could we get that as an additional option at some point?