Next week, I will be sharing strategies and tactics to better attract tech talent in a workshop with CivicAction and The Toronto Region Board of Trade. Toronto is the fastest-growing tech hub in North America with more than 81,000 tech jobs added since 2016. A few months ago, we highlighted the competitive recruiting landscape for strong tech talent in the Toronto ecosystem. Since then, we have seen the competition to land key engineering, product, and AI hires increase. In the past few months, both big tech giants and growth stage startups have announced opening additional offices in Toronto including Meta, Netflix, and TikTok, and scaling startups like Sentry. These companies recognize the extraordinary caliber of the Canadian tech talent pool.
There is also a growing awareness among prospective candidates of their value. Candidates better understand the market, with increased access to crowdsourced tools like Levels.fyi and open tech forums on apps like Blind and Fishbowl. In our ongoing work to recruit top-tier AI talent on behalf of our portfolio companies, we have seen a significant increase in salaries (external sources report as much as a 30% increase). Despite this upward trend, prospective hires are not simply pursuing higher salaries. We have seen our Toronto-based companies respond to the increased demand for local talent through improved processes, employer branding, and mission-driven recruiting strategies.
Canada has been courting tech workers for years. Unlike the United States, Canada has no cap on visas for immigrating tech workers and entrepreneurs, making it an attractive destination for researchers and software engineers who may have a hard time obtaining US visas. With recent macroeconomic shifts, it is unclear if this market will hit a ceiling. What is clear is that candidates are recognizing their position and can make choices about where they would like to live and work in the new remote-first talent landscape.
AI News This Week
-
The AI 50 2022 (Forbes)
Forbes released its fourth annual AI 50 list, which recognizes private North American companies using AI in “interesting and effective” ways. This year we congratulate Radical Ventures portfolio companies Cohere, Genesis Therapeutics, and Waabi for making the list. These companies are demonstrating the transformative power of AI across industries. Cohere is kickstarting a new chapter in responsible machine learning by giving developers and businesses access to NLP powered by the latest generation of large language models. Genesis Therapeutics provides AI-powered drug discovery and recently partnered with Genentech and Eli Lilly. Waabi is building the next generation of self-driving technology, unleashing the power of AI to bring the promise of self-driving closer to commercialization. Waabi was featured this week as part of the AI 50 series with an interview from Founder and CEO Raquel Urtasun.
-
The AI Economist: Taxation policy design via two-level deep multiagent reinforcement learning (Science)
A group of tech entrepreneurs and researchers, including You.com Founder Richard Socher, are exploring how to apply AI to economics. Their research focuses specifically on tax policy. AI-driven simulations may be able to test economic policies before enacting them, validate assumptions in policy proposals, and evaluate ideas coming from economic theory. Future research could build on these learnings and calibrate models to real-world data. The researchers will release an open-source version of their environment and sample training code for the simulation with the belief that AI-driven policy design can democratize policymaking by enabling a broad multidisciplinary audience to inspect, debate, and build future policy making frameworks.
-
Google launches a 9 exaflop cluster of Cloud TPU v4 pods into public preview (TechCrunch)
Demand for machine learning capacity, performance, and scale continues to increase at an unprecedented rate. In response, Google Cloud unveiled its new machine learning cluster powered by its Cloud TPU v4 Pods last week. Several top AI research teams, including Radical Ventures portfolio company Cohere, were granted early access to the ML infrastructure hub. Reducing the carbon footprint of ML compute has been a priority for the company. Aidan Gomez, Co-Founder and CEO at Cohere, provided a review for the product, “At Cohere, we build cutting-edge natural language processing (NLP) services, including APIs for language generation, classification, and search. These tools are built on top of a set of language models that Cohere trains from scratch on Cloud TPUs using JAX. We saw a 70% improvement in training time for our largest model when moving from Cloud TPU v3 Pods to Cloud TPU v4 Pods, allowing faster iterations for our researchers and higher quality results for our customers. The exceptionally low carbon footprint of Cloud TPU v4 Pods was another key factor for us.”
-
A quick guide to the most important AI law you’ve never heard of (MIT Technology Review)
The European Union has released the first draft of its ambitious Artificial Intelligence Act. “The mother of all AI laws,” it is the first law that aims to curb potential harms associated with AI by regulating the whole sector. If the EU succeeds, the bill would apply to 27 countries and could set a new global standard for AI oversight. The bill requires extra checks for AI use cases deemed high risk and unacceptable use cases, such as scoring people based on perceived trustworthiness, will be banned. The bill also restricts law enforcement’s ability to use facial recognition in public places. Some critics worry that the regulation will stifle innovations by creating extra red tape for AI companies. The EU counters that the AI Act will only apply to the riskiest use cases, which the European Commission estimates would apply to just 5 to 15% of all AI applications. It will be at least a year before the text is set and a few more years before businesses operating in Europe will have to comply.
-
Projects: Reinforcement Learning for real-world problems (Kurin ViTaly)
Reinforcement learning (RL) is one of the three basic machine learning paradigms, and is credited to University of Alberta Professor and Alberta Machine Intelligence Institute co-Founder, Richard Sutton. Like training a dog, the computer employs trial and error to come up with a solution to the problem and either receives rewards or penalties for the actions it performs. Its goal is to maximize the total reward. To date, RL has been chiefly effective in lab settings. But, there has been a growing number of successes from applying RL to real-life problems. Kurin ViTaly, a reinforcement learning-focused PhD student at the University of Oxford, has put together a list. The applications include energy, healthcare, finance, drug and material discovery, and entertainment. The list is not exhaustive but provides an overview of where RL is being applied today.
Radical Reads is edited by Leah Morris (Senior Director, Velocity Program, Radical Ventures).