Make a Difference, Not Just Noise

The OpenAI founder Sam Altman loves to ‘not brag’ about his team. With a team size of only 375 people, his company has been able to produce a product like ChatGPT, along with GPT-3.5, DALL-E, Codex, and MuseNet, and become a worldwide sensation.

In contrast, the rival company Google's Deepmind has more than 1000 employees. Meta AI, as per their website, has about?228 researchers?across research areas.?

Experts say whether 50 or 300 or 1000 employees, it’s not justified to make such comparisons between companies. Each enterprise is involved in solving a varied problem; the target group is different for each of them, and so is the nature of their undertaking.

For instance, while OpenAI is product-focused, Deepmind and Meta AI are largely research-oriented. DeepMind has produced 30 times more research papers in the past three years than OpenAI, despite having a comparatively smaller team size. Additionally, the research produced by DeepMind is more diverse in terms of subject matter. Likewise for Meta, wherein it published?104 research?papers last year.?

Some of the major inventions by Deepmind were AlphaGO, which defeated a human Go player, and the 3D protein-prediction model?AlphaFold, which solved the?50-year-old grand challenge?of protein folding. Meta also released the protein prediction model ESMFold.?

Google and Meta have played integral roles in supporting the open-source community by providing the foundation for thousands of tech startups through their research and technologies. For example, ChatGPT, developed by OpenAI, is based on the Transformer model, which was first introduced in the paper titled “Attention is all you need” six years ago by Google researchers.

While OpenAI might be making the loudest noise, its peers, Deepmind and Meta AI, are undertaking equally impactful research in the field of Artificial Intelligence.?

Read the full story?here.

Big Techs Bet Big on Healthcare

The healthcare sector has been exploring the potential use of AI for some time now, and recent developments in generative AI may further solidify this interest. A McKinsey and Harvard report suggests that using AI technologies such as machine learning and natural language processing could save healthcare costs between $200 billion and $360 billion while improving patient experience and expanding access to healthcare facilities.?

Despite initial challenges with the adoption of AI in healthcare, big tech companies have continued to invest in the field, such as Microsoft's partnership with AI healthcare startup Paige, which focuses on cancer diagnostics and pathology.

Read the full story?here.

India’s Semiconductor China+1 Dream

With the US government's new sanctions on China over the import and export of semiconductors, the world has turned its gaze towards India and whether it will be able to fill the void.

There have been some deals and developments in the past few months, such as the Vedanta–Foxconn deal and Tata’s plan to foray into the semiconductor realm. But there have been some hiccups too. Meanwhile, the country has managed to be a significant location for global semiconductor design companies as many have established research and development centres here due to the abundance of skilled semiconductor design engineers and a high number of design patents and intellectual property rights registered in India.?

Read the full story?here.

AI Watermarking

Language models, such as ChatGPT, have advanced significantly in recent years to an indistinguishable level from human writing. The University of Maryland suggests using a technique called "watermarking" as a new benchmark for assessing AI performance.?

Watermarking techniques could be used to make data generated by models detectable, though it is still in the early stages of development.

Read the full story?here.