What we have right and wrong

In December 2020, we published a list of 10 predictions about the world of artificial intelligence in 2021.

With 2021 now drawing to a close, let’s revisit these predictions to see how things go. Actually exhausted. There is a lot to learn from these retrospectives on the state and trajectory of AI today.

---

Prediction 1: Waymo and Cruise will debut in public markets.

Result: Bad

At the start of this year, no autonomous vehicle company had ever gone public. 2021 is the year that everything changed.

TuSimple, Embark, and Aurora all made their public market debuts this year. Argo is preparing to go public. Plus.ai and Pony.ai both announced SPAC deals this year (although Pony.ai has since put their plans on the back burner). Credible rumors are circulating about the upcoming public market debut for other stand-alone players.

But Waymo and Cruise are not on that list.

Given that Waymo and Cruise are the best-capitalized AV companies, it makes sense that they aren’t necessarily the first to need to tap the public markets for more capital.

Nonetheless, even though our timeline turned out to be premature, we expect both of these companies to eventually go public.

Prediction 2: A political deepfake will spread across the United States, fueling widespread confusion and misinformation.

Result: Bad

Deepfakes, which just a few years ago were an oddity on the fringes of the internet, have established themselves in public consciousness in 2021.

From a documentary by Anthony Bourdain to viral clips by Tom Cruise, from a widely doomed new porn app to a bizarre story about the vindictive mother of a cheerleader in a small American town, deepfakes are quickly a part of. of our societal environment.

But no deepfake has yet deceived a large number of viewers and caused significant real-world damage in the realm of American politics. Hopefully this will be the case in 2022.

Prediction 3: The total number of published academic research papers on federated learning will increase.

Results : .

The federated learning research activity has indeed increased this year.

The number of academic research papers published on federated learning increased from 254 in 2018 to 1,340 in 2019, to 3,940 in 2020, according to Google Scholar. In 2021, that number has risen to 9,110, with four weeks left in the year.

In last year’s forecast, we specified that this number would exceed 10,000 in 2021, hence the “ish”. This can be summed up over …

Prediction 4: One of the top AI chip startups will be acquired by a major semiconductor company for more than $ 2 billion.

Result: Bad

No multi-billion dollar acquisitions have taken place in the AI ​​chip world in 2021.

Instead, the top AI chip startups have all thrown rounds at multibillion-dollar valuations, making it clear that they don’t aspire not to be acquired but to become large, stand-alone state-owned companies.

In our forecast last December, we identified three startups in particular as likely acquisition targets. Among these: SambaNova raised a $ 670 million Series D at a valuation of $ 5 billion in April; Cerebras raised a $ 250 million Series F at a valuation of $ 4 billion last month; and Graphcore has raised $ 220 million to a valuation close to $ 3 billion amid rumors of an upcoming IPO.

Other big AI chip startups like Groq and Untether AI have also raised significant funding rounds in 2021.

Prediction 5: One of the leading AI drug discovery startups will be acquired by a major pharmaceutical company for more than $ 2 billion.

Result: Bad

In 2021, none of the top AI drug discovery startups were acquired by an incumbent pharmaceutical operator. Instead, just like the AI ​​chip startups in the previous section, these companies have raised record funds to challenge incumbents head-on.

Several players in AI drug discovery completed IPOs in 2021, making them one of the first AI companies in the world to trade on public markets.

The recursion became public in April; Exscientia followed him in October. Insilico is expected to go public soon. Insitro, XtalPi, and a handful of other AI drug discovery players have raised massive private tours this year. For most of these competitors, the acquisition window is probably over.

Prediction 6: The US federal government will make AI a real political priority for the first time.

Results :

Finally, a prediction that we nailed!

For years, US policymakers have been relatively oblivious to the strategic importance of artificial intelligence, while more forward-thinking governments like China and Canada have deployed detailed national strategies to position themselves as world leaders. of AI.

That changed dramatically in 2021, with an explosion in US public policy activity related to AI. Earlier this year, Congress passed legislation to promote and coordinate AI research. Numerous additional AI-related bills have been tabled in both houses of Congress this year. A dedicated White House group has been created to oversee the country’s overall approach to AI. The US military has gone too far in its investments in AI. In October, the Biden administration called for an “AI Bill of Rights” for the American people. The list goes on.

It would be an exaggeration to say that the U.S. government has a cohesive national AI strategy in place. But in 2021, artificial intelligence has risen to the forefront of Washington’s political agenda.

Prediction 7: An NLP model with over a trillion parameters will be built.

Results :

In January 2021, less than a month after our forecast was released, Google announced that it had trained a model with 1.6 trillion parameters, making it the largest AI model ever to be built.

The question now is: how big will these models be in 2022?

Prediction 8: The “MLOps” category will begin to experience significant market consolidation.

Results : .

The crowded MLOps landscape began to consolidate in 2021. In several cases this year, large AI platforms have acquired smaller startups creating tools and infrastructure for machine learning.

Perhaps the most notable example came in July with DataRobot’s acquisition of Algorithmia, which had raised nearly $ 40 million in venture capital.

Other examples include the acquisition of Defined AI by HPE and the acquisition of decision.ai by DataRobot.

But there has been less M&A activity in MLOps this year than expected. In last year’s forecast, we listed 14 MLOps startups that we saw as potential acquisition targets. Of these, only one – Algorithmia – ended up being acquired. (Several others on this list – Weights & Biases, Snorkel AI, OctoML – have instead thrown rounds at monster ratings.)

Prediction 9: AI will become a big part of the narrative in regulators’ antitrust efforts against big tech companies.

Results :

Regulatory momentum for antitrust action against Big Tech has been building for years given the inordinate influence that companies like Alphabet, Amazon and Facebook exert on the economy. But over the past year, antitrust regulators have increasingly refined their message by focusing on the structural advantages these giants enjoy in AI. The starting point, almost always, is with organizations’ unparalleled data assets and aggressive data accumulation practices.

From recent Senate antitrust hearings to Presidential Executive Orders, this theme of unfair advantages of data translating into unfair advantages for AI is becoming an increasingly important dimension of the Big Tech antitrust movement.

Last month, for example, Lina Khan’s Federal Trade Commission appointed prominent AI critic Meredith Whittaker to a special role as the FTC’s senior advisor on AI. As one industry observer put it: “The hiring of Whittaker is just the latest proof of the FTC’s attention to algorithms and algorithmic problems.

Prediction 10: Biology will continue to gain traction as the hottest and most transformative field to apply machine learning to.

Results :

Of the predictions on last year’s list, this was the most open and the least verifiable. Despite this, many developments in 2021 point to the continued emergence of biology as the most important and impactful of all areas of AI application.

AI is transforming drug discovery, with profound implications for the pharmaceutical industry and the future of human health. The therapies discovered by AI are now in the clinic; AI drug discovery startups are now trading on public markets.

DeepMind’s landmark work AlphaFold, which was published in July, is a testament to the almost magical potential of machine learning to uncover fundamental truths about how life works. We’ve argued in this column before that AlphaFold is the most significant achievement in AI history. As Alphabet’s big announcement on Isomorphic Labs highlighted last month, AlphaFold is just the start.

Perhaps more than any other field in AI, world-class talent and investments are flowing into computational biology. Take, for example, Eric Schmidt’s $ 150 million donation earlier this year to establish a new center at Harvard and MIT that “will catalyze a new scientific discipline at the intersection of biology and machine learning. “.

In the years to come, the application of computational and machine learning methods to biology is poised to transform society and perhaps life as we know it.

Note: The author is a partner of Radical Ventures, which is an investor in Untether AI.