Shrestha Basu Mallick
@shresbm
Followers
2K
Following
2K
Media
34
Statuses
696
Generative AI Product leader for Gemini API and AI Studio @Google; Previously @Theteamatx, @Salesforce @Docusign Opinions my own
San Francisco, CA
Joined October 2009
The Gemini 2.0 era begins with 2.0 Flash Experimental release ⚡️.📈2.0 Flash beats 1.5 Pro across factuality, reasoning, coding, math. 📳 More modalities - image and audio out (in EAP).🔧 Native tool use for Google Search, code execution and 3P functions.🆕 a new multimodal,.
We just released Gemini 2.0 Flash Experimental ⚡. Available in the Gemini API and Google AI Studio for testing, it allows developers to build interactive experiences with better performance and multimodal capabilities.
6
13
188
Gemini experimental 1206 is incredible at vision. Some pictures from a recent Costa Rica visit. Especially on the last image - I still can't see the tree frog. But my guide said it was definitely there. and so does Gemini! Now for what comes next.
7
5
109
The Gemini realtime API is kind enough to let me interrupt it to keep things moving! Thanks for all the love you all are giving it. We'll keep cooking more! . Here is Gemini working with me to edit an image in Adobe Photoshop and identify the bird at the end - a motmot.
7
8
107
Gemini API definitely growing dramatically. but we are just getting started! The goal is to be a trusted platform for every developer as they build world-changing applications!.
2/ Gemini API calls, token volume, consumer usage, business adoption - all growing dramatically. And all 7 of our products and platforms with 2B+ monthly users use Gemini models, including the newest 2B-user product, Google Maps.
4
3
101
🚂 2 new production models: 1.5-Pro-002 and 1.5-Flash-002 with improved math, vision understanding and long context performance.🧪 Improved version of experimental 1.5-Flash-8B.🛟 Updated settings on safety filters.↗️ 2x higher rate limits on 1.5 Flash and ~3x higher on 1.5 Pro.
6
5
97
You can now edit code in Google Search Generative Experience (SGE) in addition to generating and executing it!. This will let you .- play around with the generated code.- write small test cases to verify the code.- make small edits before copying into your development environment
4
12
55
Life update!.
Today I am happy to share that Google AI Studio and the Gemini Developer API (along with our teams) are moving over to @GoogleDeepMind! . This move will allow us to double down on our already deep collaboration and accelerate the research to developer pipeline. Time to build!.
5
0
47
We got lots and lots of feedback re: this. And now here it is!.
Gemini is now accessible via the OpenAI libraries! Update 3 lines of code and get started with the latest Gemini models : ).
2
2
38
Answering all your questions about Gemini 2.0 today with.@KorayKV, @JeffDean, @melvinjohnsonp and @TulseeDoshi!.Co-hosting with the one and only @OfficialLoganK!.Send us your questions and come hang out with us
6
7
36
Human raters and auto-raters often rate longer responses higer. However in reality users often prefer punchy, concise responses for non-creative use cases!. For the 2 new models we released - Gemini-1.5-Pro-002 and Gemini-1.5-Flash-002, the default output length is ~5-20%.
2
0
29
@vikhyatk Hi I am the PM on the Gemini APIs working with @OfficialLoganK . Are you using the Gemini APIs or Vertex APIs?.
3
0
27
Was great to share insights about how we built Grounding with Google Search with @labenz and @OfficialLoganK . In the craziness of launch day, it didn't register podcasts have a video mode as well - else would have made more of an effort than showing up in my partners'.
Google is dancing! 💃. Today, @GoogleAI launched "Grounding" with Google Search in the Gemini API & AI Studio 🔍. I had a chance to preview it, and I integrated Gemini into an app in <2 hours, start to finish 👏👏. @shresbm & @OfficialLoganK join to discuss
1
2
28
Incredible progress across data analysis, math. coding (we get asked about that a lot!) and reasoning (very important for whats coming!).
It is not just vibes, gemini-exp-1206 has really made significant progress (#2 overall on Livebench), can't wait to test more this weekend! 📈
2
0
25
3 new experimental models .- First release of a small model - Gemini 1.5 Flash-8B.- A Gemini 1.5 Pro model better on coding & complex prompts.- A significantly improved Gemini 1.5 Flash model. Try the model names ending with Experimental 0827
Chatbot Arena update⚡!. The latest Gemini (Pro/Flash/Flash-9b) results are now live, with over 20K community votes!. Highlights:.- New Gemini-1.5-Flash (0827) makes a huge leap, climbing from #23 to #6 overall!.- New Gemini-1.5-Pro (0827) shows strong gains in coding, math over
2
0
22
Compare mode in AI Studio out! This is one of the features that developers absolutely love in every testing !.
1
1
21
As someone commented on @OfficialLoganK 's post, this is now a Flash sale!. ⚡️ Gemini 1.5 Flash more than 70% cheaper through the Gemini API and AIS.📈 Finetuning now available to all! Tuning a model is free, inference is the same as cost as for base models.💯+ languages.
📢✨ Announcing Gemini 1.5 Flash price drop & updates! → 🔧 Enjoy the speed & efficiency of 1.5 Flash at 70%+ lower prices.⚡️ Customize Gemini 1.5 Flash with tuning.👀 Check out improvements to Google AI Studio & Gemini API. #BuildWithGemini
1
1
21
Flash 8B is in production now! Lower cost, lower latency, higher rate limits !.
2
0
18
I'll be speaking at the @AITinkerers Humans-in-the-loop Agents Hackathon in San Francisco November 2nd and 3rd. Join me and the judges:.my fellow @AITinkerers . @jpalioto @ataiiam @jamescham @jheitzeb @sara_k_48 @pk_iv @kwindla @vaibhavk97.
0
4
19
Join me at the Women in AI Summit 2024 to explore the power of Gemini APIs and Google AI Studio! We'll cover how you can start building with all the latest models and features like long context, search grounding, fine-tuning etc.
1
4
18
As a physicist, Gemini using code-execution to solve projectile motion equations makes me happy! .📸 Available for both 1.5 Pro and Flash .📸Use as a tool in the API or through AI Studio .📸 for queries that benefit from code-based reasoning such as math equations or
1
3
17
Improved spatial understanding!.
New Gemini!. I just released llm-gemini 0.6 adding support for the "gemini-exp-1206" model, and then got a pretty spectacular result for my "Generate an SVG of a pelican riding a bicycle" test prompt
1
0
16
Herrrre we go! Congrats to the Codey team! It has been an adventure! #GoogleIO #GoogleIO2023 #Codey #generativeai #bardforcode.
Announcing coding AI in GCP Vertex AI, and in the new Duet AI products!. Vertex and Duet will help you, your company, and your developers really boost their productivity when building on GCP. Powered by a family of PaLM2-derived models known as Codey ✨
0
2
15
Just a reminder to everyone! This is happening at 9.30 am PST today. Please join to find out more about what the Gemini API and AI Studio can do and hear exciting updates.
Join me at the Women in AI Summit 2024 to explore the power of Gemini APIs and Google AI Studio! We'll cover how you can start building with all the latest models and features like long context, search grounding, fine-tuning etc.
1
1
15
Back at you @OfficialLoganK . And immense gratitude for all our partner teams.
I have so much gratitude and appreciation for the AI Studio, Google Deepmind, and all of our partner teams. It truly takes a village to make these launches happen, so much fun to part of. A special group of people pushing the frontier and shipping.
1
0
16
Given the price-performance tradeoff on our recently released Flash models, we see the most potential for in tasks ranging from high volume multimodal use cases to long context summarization tasks.
1
0
14
Took @shreyas's course this weekend. Its among the top 3 courses I have taken in my career. The advice/learning is different - an unusual mix of visionary, pragmatic and actionable. But the right mix to achieve the fulfilling goals that we become PMs for.
0
2
12
Coming soon to a Colab notebook near you! .Thanks to @sammysamau @Skiminok @edtoh @zekemiller99 @popovicu94 @malmaud @MitchellAGordon @r_sonthalia for contributing to this and ofc @its_ericchu @KevinKiao @jeffistyping @_arohan_ for the model!.
Your new coding assistant is almost here! Check out these new Colab features: natural language to code generation, code completion, and an integrated chatbot. Read all about at authored by @thechrisperry and @shresbm.
1
6
13
Potential use cases we have heard for experimental 1.5 Flash 8B model ⚡️.- Initial-stage classifier or router in workflows involving larger models.- Lower latency deployment in chatbots, RAG systems.- Closer to real time video analysis .- Medical and educational simulations. We.
1
1
13
Try out Google's latest experimental model in Tell us what you think!
3
1
12
I remember the days that I would ping Bard with Bengali transcribed in English and it would respond in English. Now we have Gemini Advanced in 9 Indian languages including Bengali!.
Exciting news! 🇮🇳 Today, we're launching the Gemini mobile app in India, available in English and 9 Indian languages. We’re also adding these local languages to Gemini Advanced, plus other new features, and launching Gemini in Google Messages in English.
1
0
12
and we will keep delivering!.
Google Gemini 2.0 realtime AI is unbelievable. The fact that you can talk, share the screen, share images it's another level of interaction. What OpenAI promised, Google delivered. Amazing job 🔥
1
0
12
Happening now. Excited for all the discussion!.
0
1
12
The Gemini 2.0 multimodal live aka realtime API uses is able to connect a set of tools together including user-defined tools via compositional function calling. @alexanderchen gets movie data using search as a tool and make plots on various stats using code execution as a tool.
Data visualization + Search in real-time! 📈🔎 I was honestly really amazed at the latency when @hapticdata first got this cool demo working with Multimodal Live API. It feels almost magical to see a graph appear in near real-time on demand . .
1
1
12
Here goes! Excited to start working with @matvelloso @OfficialLoganK and @Ronenkofman on this next phase!.
The Gemini API and Google AI Studio are now available in 200+ countries, Gemini 1.5 Flash costs $0.35 per 1M tokens, with context caching coming next month. So much going on today 🤯.
0
0
11
Happy Diwali and Happy Halloween!.
Wishing everyone a bright and joyful Diwali filled with warmth and light! 🪔
3
0
11
This weekend I read the Michelangelo paper from @GoogleDeepMind ( which benchmarks models' ability to find relevant information needed for complex tasks spread across large context, much like Michelangelo chiseling away irrelevant material to reveal the
2
3
11
Using Gemini for some Friday plant care: Uploaded a pic of a plant that (shamefully) I had forgotten the name of. Gemini Pro. 🌱identified it correctly as a croton despite the bad angle on the pic (did take 8 secs to think!). 🍁told me why the leaves were not as colorful
1
0
10
AGI taking over repetitive tasks will PUSH more humans to spend more time doing complex, creative and interconnected tasks …requiring deeper thought and fewer distractions. 🗯️.That is already an IQ extender🧐.
0
3
10
Thanks! We are trying. Keep the feedback coming.
@OfficialLoganK @GoogleAI Google seems to be actually shipping more frequently now and substantial stuff too. Meanwhile, the 🍓 nonsense has resulted in nothing. I've been hard on Google over Gemini, but this is an awesome renewed sense of direction. Well done.
1
0
9
Incredible demonstration of how cheap and easy it has become to use LLMs to cut out rote work. Interestingly @simonw did not even turn on JSON mode/structured output (see right side of AI Studio screenshot) but it worked! With structured outputs we aim to make such use cases.
I needed information from a dozen emails in my inbox. so I ran a screen capture tool, clicked through each of them and got Gemini 1.5 Flash multi-modal LLM to extract (correct, I checked it) JSON data from that 35 second video. Total cost: $0.00082635
0
0
8
Very excited to have been part of this release! Thanks to both the Codey and GCP teams!.
Exciting release of this awesome product our team has been working on. Duet AI brings real-time code suggestions and a chatbot directly on your IDE, with the potential to massively increase dev productivity! #DuetAI #GoogleIO2023📷 @shresbm @Skiminok .
0
2
8
Looking forward to speaking on Building with the Gemini API and AI Studio at DevFest Silicon Valley! Come chat with us if you have feedback on the Gemini API or AI Studio.
1
1
8
Gemini 1.5 Pro experimental is #1 on LMSYS Arena! With a score of 1300!.
Exciting News from Chatbot Arena!. @GoogleDeepMind's new Gemini 1.5 Pro (Experimental 0801) has been tested in Arena for the past week, gathering over 12K community votes. For the first time, Google Gemini has claimed the #1 spot, surpassing GPT-4o/Claude-3.5 with an impressive
0
1
9
So many interesting conversations from "human-on-the-loop" to "why computers get so hot" to is "Florida the lightning capital of the US". The answer to the last one is its not. Humans also need Grounding with Google Search sometimes 😉.
Awesome AI builders dinner tonight 🤖🍽️. Thanks @jamescham & @jheitzeb for co-hosting with us 🥳. Come to our Hackathon tomorrow at @weights_biases to connect with more awesome builders 🪁. —-.CC: @altryne, @ataiiam, @tereza_tizkova,@dexhorthy, @aboutphilippe, @shresbm, @swyx,
1
1
9
Excited to see how devs build with this! Please send us feedback.
4
0
7
One of my favorite execs at Google Elizabeth Reid showcasing Search Generative Experiences now directly on Search. And lots more to come this summer!
1
1
9
Excited to be part of the panel on the voice AI meetup with the rest of these amazing folks.
Join us for the first Voice AI Meetup of the new year, Tuesday Jan 14th. I'll be moderating a panel discussion about Voice AI in 2025, featuring:. ⭐️ @krandiash — cofounder of @cartesia_ai . ⭐️ Niamh Gavin — Stanford. ⭐️ @shresbm — Generative AI Product leader for
1
0
9
Our datascience agent just launched on This agent can generate entire Colab notebooks! It helps you get started on your data science tasks by cleaning and explore your data for you, doing statistical analysis and creating predictive models. It can answer
0
0
8
Caching has a 75% discount compared to regular input tokens! Also you can try caching for Flash in the free tier upto 1MM tokens. Use caching for long context use cases where you need to reference the same context repeatedly. Some ideas we have are . 💽 Long-context ICL by.
Context caching is now available for Gemini 1.5 Pro and 1.5 Flash! 🙌. At a lower cost, context caching in the Gemini API makes working with millions of tokens a breeze. Pass content once, cache the tokens, and refer back for later requests. Learn more →
0
0
8
For all those asking how to use the Gemini realtime API with WebRTC!.
Gemini Multimodal Live + WebRTC. Build Gemini voice/video apps with WebRTC SDKs for:.- Web.- React.- React Native.- iOS.- Android.- C++. The SDKs support WebRTC, WebSocket, and HTTP network transport options. Change one line of code to switch protocols. Here's a simple Gemini +
0
0
8
Good refresher! The REFORMs checklist should be a handy tool for all ML producers releasing models in the real world.
Have you ever trained a model you thought was good, but then it failed miserably when applied to real world data? . If so, you’re in good company. Check out @michael_lones' piece on how to avoid machine learning pitfalls!.👇.
0
4
8
During my PhD, I had many friends in computational Physics and Chemistry that spent their entire PhDs discovering a single protein folding structure. Today the AlphaFold DB provides open access to over 200 million protein structure predictions and has enabled advances in.
This year’s chemistry laureates Demis Hassabis and John Jumper have developed an AI model, AlphaFold2, to solve a 50-year-old problem: predicting proteins’ complex structures. Check out two examples of protein structures determined using AlphaFold2. First up, a bacterial enzyme
1
0
8
Thanks to @jheitzeb @kwindla and the rest of the @trydaily folks for making this happen! Showcasing the Gemini Multimodal Live capabilities with the @pipecat_ai repo! You can see how fast it is, how graciously interruptible. And it identified the guitar!.Someday we might even.
Correctly identifying guitar sounds with Google Gemini 2 voice-to-voice
0
0
8
Top of the leaderboard! #1 on vision, math, hard prompts and creative writing! Congrats to the team!.
Massive News from Chatbot Arena🔥. @GoogleDeepMind's latest Gemini (Exp 1114), tested with 6K+ community votes over the past week, now ranks joint #1 overall with an impressive 40+ score leap — matching 4o-latest in and surpassing o1-preview! It also claims #1 on Vision
0
1
8
Easy way to get started with our Multimodal Live API.
Jump right into the Multimodal Live API! This notebook provides a baseline introduction to the Gemini 2.0 Multimodal Live API. ↓.
0
1
7
My favorite use case for Compare mode is to see how the model performs without and with advanced settings. Here you can see how the model solves the same chemistry problem without and with code-execution-as-a-tool turned on!
1
0
7
Thank you for changing your stance. Let us know what else we can do for you 😍.
Gemini 2.0 is just insane. They really did surprise me well. And my stance towards them has completely changed.
0
0
7
Yes with both screensharing and camera input!.
@sundarpichai Gemini Flash is truly multimodal, check out this video! Reads a creditcard statement , identifies the fruits, basically no lag! We are deploying this today at !
0
0
7
We want to add support for meaningful use cases to make the flow truly frictionless. So keep the feedback coming.
Google Gemini 2.0 realtime AI is insane. Watch me turn it into a live code tutor just by sharing my screen and talking to it. We’re living in future. I’m speechless.
0
0
7
We love hearing from devs! Positive feedback but especially all the ways in which the model and APIs can do better. Come meet the team at
1
0
6
PaLM-E helping robot fetch snacks from the microkitchen. Love it! Can my robot beat your robot to get that last bag of Miss Vickies?.
0
2
6
Very very excited for this! Still remember the early days of brainstorming when this was just thoughts on a Google doc. So proud of the team and the friends I have made along the way!.
Announcing General Availability of Duet AI for Developers and Duet AI in Security Operations @googlecloud
1
0
6
New Udacity course using the Gemini API and Google AI Studio! Developed by Berkeley and Google experts!.Try it and let us know what you think!. You will learn how to build an app using the Gemini APIs and how to use AIS to design better prompts!.
🌟 Calling all developers – our new Gemini API by Google Course is live! . This FREE course is your gateway to mastering the integration of generative AI into your applications, websites, and services. @googledevs . Enroll in this free course today:
0
0
6
Great to see the Gemini models performing so well! Especially the validation of the outstanding price-performance for 1.5 Flash!.
Big news – Gemini 1.5 Flash, Pro and Advanced results are out!🔥. - Gemini 1.5 Pro/Advanced at #2, closing in on GPT-4o.- Gemini 1.5 Flash at #9, outperforming Llama-3-70b and nearly reaching GPT-4-0125 (!). Pro is significantly stronger than its April version. Flash’s cost,
0
0
6
Been trying to make this happen for months! So happy to finally be in the Western Balkans (Albania coming soon!).
Today we are expanding access for the Gemini API and Google AI Studio to the following countries: . - Bosnia and Herzegovina.- Montenegro.- North Macedonia.- Serbia.- Ukraine. Thanks for the patience while we made this happen (Albania coming soon)! 🌍.
0
0
6
Thanks for the great response to the features we shipped yesterday! Excited to have partnered with @OfficialLoganK @Ronenkofman @matvelloso and others on this launch! More to come very soon!. 🚀 Gemma 2 now also in AI Studio.🚀 Code-execution as a tool through API and AIS.🚀.
0
1
6
And using Search as a tool on the multimodal live experiences has been unlocking a lot more high quality, factual use cases!.
Throughout this year, we have had a razor focus on improving the factual accuracy of Gemini models' responses in various scenarios. This result is on the Vectara hallucination leaderboard for the Gemini 2.0 Flash model that launched today:
0
0
6
Catching on tweets! Love this demo where the realtime API works with Search grounding to identify that different sources count total number of West Wing episodes differently depending on whether specials are included.
p.s. As a West Wing fan, my favorite part is how the model acknowledged that you could either say there are 154 or 156 episodes (it depends on if you count the specials!)🙂 Fuzzy data retrieval with explanations is really cool.
0
0
5
@googledevs @koraykv @JeffDean @melvinjohnsonp @tulseedoshi @OfficialLoganK Question: How does the latency of Flash 2.0 compare to Flash 1.5?.Response: Similar but of course depends where in the stack you are testing. We also expect latency to improve over time.
1
0
5
Check out the recent #DuetAI blog post written by @grappeggia and me. If you are a company evaluating #GenerativeAI tools for your enterprise needs, we have a checklist to guide you!. #googleai #Google.
0
2
5
Wanted to extract MIME subtype from a list copied from Gemini API docs. Once upon a time this would be copy/paste and then laboriously delete unwanted characters. Now its just a prompt!
0
0
5
And so it begins! Excited for all that will be unveiled today. As Gemini says."In the depths of the code, a human heart beats,.Each product a monument, which vision completes.Its journey from thought, to tangible thing,.A testament to progress, a song we will sing."
1
0
5
@OfficialLoganK @sithamet While that does not address all the issues on this thread, we are making some updates to the civic integrity filter soon.
0
0
5
Excited to start using the Gemini models for code AI products @Google! . Gemini Ultra advances SOTA in 30 of 32 benchmarks including HumanEval (74.4%) and Natural2code (74.9%). Gemini Nano will unlock lots of on-device possibilties.
Seeing some qs on what Gemini *is* (beyond the zodiac :). Best way to understand Gemini’s underlying amazing capabilities is to see them in action, take a look ⬇️
0
0
5
Streaming mode is pure science fiction!.
Wrote up my initial impressions of the new Google Gemini 2.0 Flash model - it's really good, and the streaming mode (where you can stream video and audio to it and get audio streamed right back) is pure science-fiction
0
0
5
Very proud moment!.
Today we are launching Data Science Agent, a Google Labs experiment. ◆ Helps with complex tasks like planning, and error correction. ◆ Helps with data science tasks like predictive modeling. ◆ Outputs an AI-generated Colab notebook based on your prompt. Try it at
0
0
5
We wish we thought of this phrase first! Thanks @DamdiPawlowski.
@OfficialLoganK @GoogleAI flash sale.
1
0
5
So excited for this demo series!.
0
0
5
Have had lots of folks ask about bounding boxes recently!.
Thanks to its ability to detect bounding boxes, Gemini is incredible at coding UIs from images. I built an agentic system that looks at a UI, finds the bounding boxes, codes it, and refines the results based on the original image. 🧑💻. The best results I got from any model!
0
2
5
Excited for coding updates in SGE!. Congrats to @sammysamau @malmaud and other members of the Codey and SGE team. #googlesearch #googleai #SGE #GenerativeAI #LLM.
New updates to our Search Generative Experience include definitions with related images to help explain complex topics, more coding capabilities, and a new experiment that helps you find what you're looking for in a long article more easily.
0
0
5
We will take it as feedback @emollick and Magic to the eval set 😉.
Sadly, Gemini 2.0 Flash cannot quite play Magic the Gathering. It identifies the game and strategy, but fumbles the cards, and sometimes has vision issues. MtG benchmark is currently undefeated. (Sound on if you want to hear what the AI is saying)
0
0
4
Make up your mind Google Lens! If you must know, I prefer the bottom one
0
0
3
Next up @matvelloso doing this in our native languages. hopefully soon!.
A quick demo of Google AI Studio screen sharing capability with Gemini 2.0
0
0
4
Such an informative post by @kwindla! Detecting when to respond in a conversation is something we do so intuitively as humans. But while building the realtime API we realized it’s so difficult to get it right. Lots of improvement still to come.
Better/faster/cheaper voice AI turn detection with Gemini 2.0. The code that determines when the agent should respond to the user is some of the most important code in your voice AI agent. The technical terms for this job are "turn detection" or "phrase endpointing.". If the
0
0
4
@Ypermythos @vikhyatk @OfficialLoganK thank you! We wish we could be more responsive. Enabling devs to build meaningful apps with our APIs would be the biggest reward for all our efforts.
0
0
4
2 of my favorite people talking all things Gemini! Check it out!.
.@tulseedoshi (Gemini model product lead) joins @OfficialLoganK to go behind the scenes of Gemini 2.0, taking a deep dive into the model's multimodal capabilities and native tool use, and our approach to shipping experimental models. ↓
0
0
4
2nd most expensive ticket of my life. Coming to the airport at 3 am to find it packed. But a trip with the girls - priceless.
0
0
4
Apply to be a KaggleX Fellowship advisor!.
Share your knowledge and make an impact with early-career AI/ML practitioners by applying to be a KaggleX Fellowship advisor! We're looking for more advisors to apply before July 21st.
0
0
4
Product thought of the day: There's lots of folks building "0-to-1". Now the focus needs to shift to "prototype-to- product-market-fit".
0
1
4
And one more thing 🙂 The same PaLM2 / Codey technology is now powering Android Studio Bot - generating code, answering questions, and helping with unit tests right there in your Android Studio IDE! ✨
0
1
4
We have addressed cost! Latency improvements soon!.
I've been raving about caching being underhyped since IO, most of the long context nay sayers point at price and speed as the two things of why long context will never come close to their precious RAG solutions. Context caching from google is an answer to both these points!.
0
0
4
Get help with your coding queries . now in India! . Congrats @pottimouth42 @sammysamau @malmaud @MitchellAGordon and to the rest of the Labs and Search teams.
0
1
4
Many great insights in this podcast. Some that stood out: .1) A big pursuit in 2024 will be finding sustainable AI use cases. Looking in non-tech companies outside of Silicon Valley can yield good ones.2) Developers don't want "the solution"- they want intuitive, reusable.
"We think that practically every internal app is going to be AI infused over the next three years." Hear from our founder @dvdhsu on the future of internal tooling 👇.
1
1
4
We could use Gen AI to automate the work needed to cut through bureaucratic red tape to get approvals. But then that Gen AI will need to get approved ….
India just kissed its future goodbye! . Every company deploying a GenAI model now requires approval from the Indian government! . That is, you now need approval for merely deploying a 7b open source model 🤯🤯. If you know the Indian government, you know this will a huge drag!
1
1
3
@altryne @googleaistudio Almost. Its 10 RPM just for the experimental phase. We are obviously always looking into adding limits.
1
0
4