On MIT’s Technology Review, the pioneer of AI, Geoffrey Hinton, stated that deep learning “will do everything.” At least eventually. We are still waiting for some breakthroughs for the technology to reach the status, however.
The history of deep learning actually reaches the 1980s. The problem then was a simple one; there was not enough power for the approach to work. 2012 was the year a true breakthrough happen. During a competition to build “computer vision systems that would recognize 1,000 objects.” All teams struggled to achieve 75% accuracy. Then Hinton’s team came, and blew everybody away. They won by an impressive 10.8 percentage points.
Now, in the interview for the MIT’s publication, he reaffirmed, that deep learning will be enough to “replace all of human intelligence.” As we mentioned, a couple more breakthrough are needed to reach the goal. One of them was achieved by Ashish Vaswani et al. in 2017. The team’s research helped interpret word meanings. There is an issue with the scale of things, however.
The human brain has about 100 trillion parameters, or synapses. What we now call a really big model, like GPT-3, has 175 billion. It’s a thousand times smaller than the brain. GPT-3 can now generate pretty plausible-looking text, and it’s still tiny compared to the brain.
Most likely, the research will still take so much time, we don’t have to occupy our heads with the vision of human-like AI.
Niantic, the company famous for making children go outside for once, is now on a journey to develop their metaverse. To accomplish that, they have just raised 300 million dollars at a $9 billion valuation, TechCrunch reports.
The company known for their Pokémon Go is another company to focus on creating immersive virtual worlds. Other companies are, e.g., Meta, or Microsoft. One has to remember, that the Pokémon AR game already focus on creating a virtual world. The company now creates a 3D map of the world with their Lightship Augmented Reality Development Kit. Crucially, it seems as if the focus is on the integration between the real world, and the virtual one. It is not about replacing one with the other, as of now, at least.
Python, one of the most popular programming languages in the world, received a new major update. The new release, 3.10, brings some much-needed improvements, such as:
One module, distutils, got deprecated, so keep an eye out for that.
For the full list, read the announcement.
Thanksgiving is a celebration of the things we have. It is also time one spends with family, and close ones. As such, it should come as no shock, that during this time, people in the US stopped shopping, and browsing the internet for a while.
In their post, Cloudflare, has shared the scope of the transition. Starting with the Thanksgiving Dinner at 4pm on November 25th, the traffic started to dip solidly, remaining lower, until it spikes back up on Saturday. Cyber Monday is another spike, albeit lower than the Saturday one. Black Friday discounts were present all Thanksgiving week, however, solidly stress-testing retailers systems.