Siddhartha Agarwal @sagarwal

The layer that can route to the best AI model for the particular job is going to increase in value substantially. There are at least 3 big reasons: * Cost optimization: there are plenty of use cases where you need frontier intelligence for some tasks and something far cheaper for others. Even in the same task you may use frontier intelligence for planning and review of the work, but an OSS or cheaper model for the bulk of the workload. This is going to be standard across large buckets of work going forward. * Capability maximization: despite the bitter lesson and models generally getting better in the same direction, there are still lots of differences between models. Some are better at tool use, others better at coding, and others again better at certain domains of knowledge work. The ability to route between these at different times is a huge advantage. * Risk mitigation: while the Fable situation is somewhat of a black swan, it’s possible we’re heading toward a regulatory environment where governments may restrict models at different times based on their approval mechanisms or new things they discover. This means you’re going to want flexibility in being able to deploy workloads across different providers as a form of risk mitigation. Ultimately, it’s going to increasingly be a a strategic advantage for the applied AI layer that they can effectively route between models. Will be very interesting to see how this evolves.

OpenRouter @OpenRouter

3 weeks ago

Introducing the Fusion API, the smartest compound model in the market. Fusion achieves Fable-level intelligence at half the price. How it works 👇

723 2K 15K 6.3M 14K

“Surprise & Delight” @Delta … landed in Seattle, Jared at plane door with a “Agarwal Family” sign, taken down to tarmac, Porsche whisks away to gorgeous Delta Sky Lounge!! Love personalized service, am huge fan, now family fans too❤️😀 Status=💎

keyboard_arrow_left Previous keyboard_arrow_right Next

@chefviveksingh @hotdinners Vivek, you were amazing in taking care of us with a special jumbo prawn dish at our 20th anniversary… we’re flying from SFO to London for our 25th on Dec 3rd … making a reservation again, hope to see you there!

keyboard_arrow_left Previous keyboard_arrow_right Next

Our incident response team is actively monitoring the war in Ukraine and working around the clock, focusing on the safety and security of our employees, users, and customers.

Accelerating partner, customer success with doubled spend in #GoogleCloud partner ecosystem; expanded regional teams for ISVs, tech resources to help move ISVs to #SaaS models, monetization at edge cloud.google.com/blog/topics/pa…

Cheers to 23 years! Thank you to all of you who inspire and motivate us to continue building a more helpful Google for everyone. Excited for what's to come...but first, time for some birthday cake! 🎂🎉

A trip down memory lane...exactly 23 years ago Google signed its first datacenter contract.  Let's walk through the lease in a thread. First, the copy you see here was sent by a "fax machine" which was something people used back then.

Together with @Workday, we will enable businesses to accelerate their digital transformations and run core Workday applications across healthcare, financial services, and retail on @googlecloud. zdnet.com/article/google…

I'm optimistic that @workday's customers will enjoy lower latency and improved scalability as part of this new partnership with @googlecloud. zdnet.com/article/google…

23 leaders, 1 goal. We’re thrilled to introduce @CapitalG’s 2021 go-to-market advisory board—industry experts from across @Google ready to help our portfolio companies scale at pace. #TeamCapitalG Meet the advisors: goo.gle/GTM

With the new @GoogleCloud powered virtual agent, it is easier for people around the world to access the COVID-19 vaccine and schedule appointments. computerweekly.com/news/252499352…

Our expansion into Chile, Germany, and Saudi Arabia brings our reach to 24 regions and 73 zones in 17 countries! Proud that @googlecloud is operating these expansions committed to sustainability and focused on renewable energy 🌎 forbes.com/sites/ilkerkok…

#Anaplan, #GoogleCloud partnership delivering intelligent, performant innovation for enterprise planning; customers access deployments globally in-region, meet local data residency/compliance requirements, blend 1st party, 3rd party data, leverage AI/ML zdnet.com/article/anapla…

Engaging discussion with @guyhellier on how #OpenText is using BigQuery, Vision AI, and Google Marketing Platform to deliver a new Digital Experience SaaS solution; how ISVs are co-innovating with #GoogleCloud, broader Google PAs cloud.withgoogle.com/next/sf/sessio…

Google and #Amwell partner to transform patient and clinician #telehealth experiences leveraging unique AI/ML and collaboration tools, and make the healthcare experience simpler, more accessible for underserved populations; includ…lnkd.in/dcNpcei lnkd.in/d3cxYRU

New @GoogleCloud Center Of Excellence Provides ISV/SaaS ‘Evangelism’ bit.ly/2DD3GVP @CRN @GCPcloud @gcloudpartners @sagarwal @Temenos @OpenText #tech #technology #IT #cloud #cloudcomputing #Google #GoogleCloud #GCP #ISVs #SaaS #software #AI #artificialintelligence #ML

Launching #GoogleCloud's ISV/SaaS CoE to accelerate cloud journeys for ISV and SaaS partners (@Temenos, @Opentext): leverage multi-cloud architectures, AI/ML, voice, intelligent insights, or co-innovate with Google Shopping, Ads, Pay and Chrome. cloud.google.com/blog/topics/pa…

Here is data on instance time of billable containers: 8000 request/second with Avg latency of 700ms for a spend of approx 500 euros/month

MediaMarktSaturn is a consumer electronics retailer that moved to #containers, #serverless and went live in 5 months using CloudRun; now focused on Analytics with Kafka (fetching data with cloudrun), BigQuery, DataStudio @googlecloud #GoogleNext19

keyboard_arrow_left Previous keyboard_arrow_right Next

Cloud TPU pod scalability: giving one from 16 chips to over 1000 chips with near linear scalability (flexibility of what portion of pod used); Recursion pharma went from 24 hours to train models on a local GPU cluster to 15 minutes @googlecloud #GoogleNext19