The social network known as Gab is back online. Its previous host—GoDaddy—withdrew support and hosting because of accusations that Gab was a haven for alt-right users. The Pittsburgh synagogue shooter was a member.

Rob Monster, the founder and CEO of Gab’s new host Epik, wrote a blog post defending his firm’s decision. He acknowledged that Gab had a duty to “monitor and lightly curate, keeping content within the bounds of the law”.

However, he also said he felt that Gab’s founder Andrew Torba was “doing something that looks useful” and described Gab’s removal from the internet as “digital censorship”.

I’ve never used this platform but if it really was mostly comprised of alt-right people, therein lies the problem. It becomes an echo chamber that only serves to reinforce pre-existing opinions and world views. Even if it was a left-leaning platform, the echoing results would likely be similar.

Google has made a particular machine learning technique open-source. It’s called Bidirectional Encoder Representations from Transformers (BERT), used for natural language processing (NLP).

With this release, anyone in the world can train their own state-of-the-art question answering system (or a variety of other models) in about 30 minutes on a single Cloud TPU, or in a few hours using a single GPU.

For more technical details Google released a white paper called BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding.

Researchers from McMaster University set out to examine smartphone environmental impact. Their study shows that buying a new phone consumes as much energy as using a phone for ten years.

Smartphones are particularly insidious for a few reasons. With a two-year average life cycle, they’re more or less disposable. The problem is that building a new smartphone–and specifically, mining the rare materials inside them–represents 85% to 95% of the device’s total CO2 emissions for two years. That means buying one new phone takes as much energy as recharging and operating a smartphone for an entire decade.