Google Unleashes Generative AI Without A Moat to Defend

Google?recently unleashed a?deluge of AIs. At the?Google I/O?keynote, the chief,?Sundar Pichai, incessantly uttered AI – approximately 27 times – AI here, AI there, and AI everywhere. However, there is a twist, almost everything showcased came with a ‘coming soon’ or an ‘experimental’ clause – aka?Labs. This might have been one of the?smartest moves?ever?after the Paris debacle?– unleashing generative AI without a moat to defend.

Even though?Google I/O?was executed smoothly to an extent, it was still a disappointment for many developers and users. There were high expectations, especially regarding integrating more powerful models to enhance capabilities, adding conversational AI features to the flagship search engine, and introducing a personalised Bard to Google Assistant, among other anticipated updates. Regrettably, none of these announcements materialised or stood out.

However, Google was?deeply invested in PaLM 2?– the GPT-3 and GPT-3.5 equivalent or even surpasses them (who knows?). Google stated that PaLM 2 is their latest infrastructure and it is used in fundamental research. “We are announcing over 25 products and features powered by PaLM 2 today,” remarked Pichai, saying that it delivers excellent foundational capabilities across a wide range of sizes, namely Gecko, Otter, Bison and Unicorn.?

Some of the products powered by PaLM 2 include?Bard,?Workspace,?Med-PaLM 2,?Vertex AI,?Duet AI for Google Cloud?and others. Now that?Bard?is available in India, we tried it last week, and it is nothing close to?ChatGPT?or?Bing Chat.?

Also, there were very few announcements around?Project Gemini, which is the equivalent of OpenAI’s GPT-4 and beyond, backed by?Google DeepMind. “While still early, we are already seeing impressive multimodal capabilities not seen in prior models. Once fine-tuned and rigorously tested for safety, Gemini will be available at various sizes and capabilities, just like PaLM 2,” he added.?

That’s not all. There was a very minimal emphasis on?Google Assistant?and?LaMDA, which had been the hot topics of discussion at Google I/O for a long time. During the event, the team did introduce LaMDA 2 as a part of the?AI Test Kitchen, but it lacked significant impact. Regarding?Google Assistant, there were a few announcements highlighting the efforts to enhance the conversational experience for users, but nothing groundbreaking.

For now, Google’s ‘bold and responsible’ approach to generative AI is the only moat the company is currently chasing, all in the hope that?developers and users will someday perceive?meaning out of its AI-powered products and offerings.?

Read:?Google-Backed Startups Are Using OpenAI’s GPT, Should It be Worried?

Top Stories of the Week >>

How IBM’s Watson X is Scripting the End of Business As We Know it

When Watson first arrived, it was supposed to change everything. But over the course of a decade, IBM’s artificial intelligence couldn’t achieve its grand vision to remake industries or generate riches for companies. As it turns out, the problems Watson hoped to solve, from tackling cancer to climate change, were?much harder than anticipated.?

This time around IBM has resisted the urge to make and invest in overarching claims and instead focused on more narrow use cases. See how IBM is carving a niche in the ever-evolving landscape of generative AI?here.?

ChatGPT’s Code Interpreter May Make Data Scientists Obsolete

OpenAI recently launched the?code interpreter plugin, which has already sparked concern among data scientists with its mere glimpse. This powerful plugin has the potential to replace numerous essential workflows for data scientists, such as visualisation, trend analysis, and even data transformation.?

When considering this code interpreter in conjunction with other remarkable advancements in the field of data science, one can't help but ponder: are data scientists on the brink of obsolescence??Read to find out.

Is the Big Data Lake Era Fading?

Big Data lakes offer advantages over traditional data handling approaches, serving as centralised repositories for all data. However, as the technology matures, challenges arise. Improper management can often lead to data lakes becoming unmanageable data swamps, with complexity, data quality, and security concerns, among others.?

Despite these challenges, data lakes provide scalability, cost savings, and functionality. To optimise benefits, a hybrid approach is emerging, with smaller distributed data lakes replacing monolithic ones.?

Learn more about this approach here.

AIM Videos >>

Exploring Remote Work Innovation

In the latest episode of the Imagination Sessions, AIM gets in touch with?Madhusudhan Krishnapuram, the VP and managing director of GoTo, a remote work solution provider, to discuss the rising demands for their services, alongside the implementation of generative AI tools and products into their offerings and more.

AIM Events >>

Attention all Generative AI enthusiasts! Join?Rakuten India, the Bangalore-based tech hub for Rakuten Group, Inc, in partnership with AIM, for the?Rakuten Product Conference?(RPC) 2023. This virtual conference will take place on?May 31st and June 1st, 2023. Don't miss out!

Register now!

AIM Shots >>

Meta recently?introduced?the?AI Sandbox, a testing playground for early versions of new tools and features, including generative AI-powered ad tools.?

In a first-of-its-kind move?towards AI adoption,?Zerodha?recently launched an internal ‘AI policy’ to address job loss concerns.?

The enterprise application software vendor?SAP?recently introduced an end-to-end data cloud in collaboration with?Google Cloud?that brings data from across the enterprise landscape using the SAP? Datasphere solution. Is this the next big thing in business intelligence??Read to find out.?

Allen Institute for AI (A12) announced?AI2 OLMo, an open, generative language model, comparable in scale to other SOTA large language models at 70 billion parameters. Learn more about this model?here.?

CRED recently upgraded its UI design with the launch of a new design experience called?Charcoal, an addition to their earlier?design languages?consisting of Topaz, Fabrik, Copper and?NeoPOP.?

Hugging Face?(HF) recently released?Transformers Agent?which allows users to manage over 100,000 HF models by conversing with the Transformers and Diffusers interface. In addition, it also released a?new open-source language model for code?called?StarCoder, in collaboration with ServiceNow.?

After two years, GitHub has finally released its updated code search engine.?Read more here.?

After the highly anticipated?release of GPT-4, OpenAI has secretly released?GPT-4-32k?API, as confirmed by several developers who have signed up for the waitlist.??Read more here.