6 ms·
Justin Trudeau announces $2.4B for AI investments in Canada
- andy99 3y agoCanadian government "investments" like this end up just sucking all the oxygen out of the ecosystem and incentivizing everyone to work to get government money instead of actually working on something industrially relevant. This will be a net loss for the country and a gain for a few well connected groups.
- belter 3y agoIt's exactly the same with EU Funds.
- svara 3y agoThat's a risk, but really in Europe the alternative often isn't to find private investment somewhere else, but to continue working for some big company or university or whatever. The EU runs the programs it does because otherwise no one would be funding innovation.
- wongarsu 3y agoMany of the big EU grants are very broad, so you can bring nearly any project and argue your case. Sure, certain things increase your chance of acceptance, and those don't always align with industry needs, but the same can be said about any funding source. It's not like VCs are known for wisely allocating money according to industry needs either. What matters is giving enough projects a chance and letting the market solve the rest. EU grants do have other issues and are difficult as your sole source of money. But I don't think funneling everyone away from industry needs is one of them.
- mistrial9 3y agothis is overly pessimistic -- for example 'Horizon 2020' in the EU had half failures perhaps.. but a couple of the projects are really strong and now growing.. so, no.
- postepowanieadm 3y agoThat's really optimistic.
- abdullahkhalids 3y ago> The bulk of that — $2 billion — is going into a fund that will aim to provide access to computing capabilities and technical infrastructure. What else should the Canadian government do? How do other more successful countries use such money to jumpstart some industrial sector?
- barfard 3y agotax cuts instead of subsidies might do better
- abdullahkhalids 3y agoInterestingly, last year, Canada introduced tax credits to support the clean energy technology sector [1]. Its "total over $60 billion over the coming ten years" [1]. But there is also some actual money in the same package. So I wonder, what the thinking of the government is in deciding when tax credits and when direct subsidy. [1] https://www.canada.ca/en/environment-climate-change/news/2023/04/minister-guilbeault-highlights-the-big-five-new-clean-investment-tax-credits-in-budget-2023-to-support-sustainable-made-in-canada-clean-economy.html https://www.canada.ca/en/environment-climate-change/news/202...
- depereo 3y agoTax cuts are used more often when the government wants to see more private investment into an industry. Not really an issue with 'AI' right now. Direct subsidy is more common when the government wants to see a particular outcome; then they can more easily steer those funds to realize it.
- 1attice 3y agocitation needed.
- mitthrowaway2 3y agoI would say not necessarily cuts, but at least tax reforms to de-risk equity compensation. Canada doesn't have an 83b election equivalent, and it's easy to end up losing a lot of money through taxes on illiquid stock options that end up amounting to more than the stock options are eventually worth. An alternative, for example, would be to allow paying taxes on stock options "in-kind": have the option to give the government a share of the stock options themselves, rather than paying that percentage in dollars on a valuation that you can't actually sell them for.
- ein0p 3y agoDrop the word “Canadian” - it applies worldwide, with very few exceptions.
- eli_gottlieb 3y agoUmmm... where do you think industrially relevant research comes from?
- 1attice 3y agoThat's the crux of the problem. 1. You have a country that is, broadly speaking, industrially oriented towards selling raw goods (wood, metals) to other countries, from whom you buy back finished products. It might not technically be mercantilism, but the dynamic is the same, and, like mercantilism, it's a mug's game. It's also increasingly unpopular with your population, which is increasingly aligned with environmentalists, because of this whole world-endy forest-firey junket that's on trend. (Fun aside: look up 'TMX'. They will never get that built.) 2. So: you want to move to a more modern economy that doesn't make you a vassal state to 'allies' (who are also your customers, who are also your employers, who also run the general store from which you buy your shovels,) and does not inflame the ire of your increasingly restive voters, who have just spent a summer inhaling the fumorous remains of several hundred thousand trees. (And not the fun kind.) So, why not just laissez the faire and let the economy sort it out? 3. (Answering 2) Because, it turns out, that unguided markets entrench existing patterns. Left to their own devices, I'm sure Canada's engineering firms would come up with _even better ways_ of ripping oil from tar sands, but, again, by 1. and 2., you really don't want to do that these days, and besides, what are you going to do when demand for dilbit goes down? There are these 'solar' panels, now, you see, and --- And so, 4. You meddle with the shape and direction of the market. You know this will cause ineffiencies, and you know that some quotient of well-connected asshats will get rich from this by doing essentially nothing. But it's still more appealing than sticking with 1 and 2. 5. If you can do this in a way that hooks into existing trends (and AI is, for most normal people, still a Big New Thing,) you may also win a leapfrog opportunity over your neighbours, pulling a TSMC-but-AI out of thin air. (Perhaps the aurora borealis.) That's a lotto ticket worth scratching. Just ask Taiwan. That's why the government funds things up here. Because they wouldn't happen at all, otherwise, and that outcome is less optimal than having them happen inefficiently. The waste along the way is a necessary evil. No pump is 100% efficient, but if gets water from the well, who cares? In a hundred years, we don't want to be still selling maple syrup to the ever-warring states of Trumpida and Greater Oregonifornia. We want to be selling the shovels. (A nation can dream, can't it?) And this fund, which is less than a 40th of what was spent on a nonexistent and unworkable oil pipeline, helps get us closer to that.
- skeeter2020 3y agoWe are huge exporters, but I don't agree with your premise; one of the things we export in huge volume is the engineering talent needed to make progress. These are young economic migrants, who leave to enrich a new country, typically the US. If they come back it's generally after they've made their career contributions and are ready to draw on a more generous social system. >> You meddle with the shape and direction of the market. It's worse than that; they're picking winners and losers, and that's inexcusable.
- WhereIsTheTruth 3y agoNuclear energy is the fruit of government money, what VCs do with AI currently is a waste of both time, energy and talents, generative AI is not what's important https://www.ft.com/content/0cdf78f9-e2cc-48ff-ba65-027f1cf83334 https://www.ft.com/content/0cdf78f9-e2cc-48ff-ba65-027f1cf83...
- WhereIsTheTruth 3y agoObviously downvoted, hence why the current climate ;)
- throwaway5959 3y agoSo what Microsoft is going to spend in a couple of months?
- resource_waste 3y agoAren't we finding out you can get ChatGPT3.5 quality models for ~1-10M dollars? Not sure how much it costs to make chatGPT4 quality, isnt it just including human reinforcement? Maybe that will get expensive.
- lupusreal 3y agoMost of this money will probably end up in Microsoft's wallet anyway, buying compute from them.
- darepublic 3y agoRemember arrivecan. Canada's attempt to make a crud app
- animex 3y agoThis is only enough to make 40 crud apps based on the ArriveCan expenditure :(
- speed_spread 3y agoNothing to do with AI or this announcement. But yes, public spending in technology is most often very inefficient.
- morkalork 3y agoWhen the consultants bidding on the contract also work for the government and are skimming off the top before subcontracting it out, then yes, that does tend to happen.
- skeeter2020 3y agoand no one, consultant or bureaucrat is to blame, with the bidding companies on to new projects, and the senior public servants all conveniently in new roles & portfolios now that it comes to light.
- barfard 3y ago[flagged]
- gnabgib 3y agoWhile CBC might be a better source, there's already discussion:[0] (24 points, 2 hours ago, 29 comments) [0]: https://news.ycombinator.com/item?id=39962194 https://news.ycombinator.com/item?id=39962194
- morkalork 3y agoWhat happened to Element AI? That was $100M up in smoke.. Also, the government is pushing some kinda harsh AI regulations at the same time. It's kinda counter intuitive.
- amitdeshwar 3y agoBought by ServiceNow. https://www.servicenow.com/company/media/press-room/servicenow-to-acquire-element-ai.html https://www.servicenow.com/company/media/press-room/servicen...
- morkalork 3y agoYes, I know. What I mean is: How did the AI darling of the country, headed by one of the biggest names in AI, with a $100M warchest totally blow it and up getting pawned off to Service Now?
- amitdeshwar 3y agoI'm not an insider, but as far as I can tell they never had a vision or a strategy, so ended up just doing "AI" consulting for random enterprises. Surprise, surprise that wasn't a path to a successful business.
- morkalork 3y agoI guarantee you the exact same insiders are lining up for, and will get first dibs, on the announced investments. After all, they've got industry experience!
- strongpigeon 3y agoThis reminded me of Element AI as well. I happen to know someone who worked there for a while who said that the people there seemed much more interested in publishing press releases than creating sellable products…
- speed_spread 3y agoIt would be fun to see Matrox get back in the GPGPU / NPU game with this.
- snek_case 3y agoDoes Matrox still make and sell GPUs? The website suggests they still make chips but from what I can see it looks like they're maybe focused on video ASICs? I would assume DSPs or something like that?
- mrkramer 3y agoMeanwhile down south Microsoft casually pumps $10bn+ into OpenAI.
- ChrisArchitect 3y ago[dupe] Some more discussion: https://news.ycombinator.com/item?id=39962194 https://news.ycombinator.com/item?id=39962194
- yieldcrv 3y agoI think it’s really funny how many industries are just state supported exclusively, while the US keeps cranking out these self organized powerhouses that cater to a market Permission to fail is a big component of that in the US
- strongpigeon 3y agoWhat I fear about investments like this is that you end up with companies that become really good at optimizing for getting grants without actually creating anything of value. Thinking specifically of Element AI.
- skeeter2020 3y agoThis is the fate of most of our largest charitable foundations. They are far better at raising funds than curing their namesake cause. The United Way is a prime example that also siphons off a sizeable percentage along the way, much like our governments.
- pknerd 3y agoumm.. when is the election in Canada?
- perardi 3y ago“Trudeau made the announcement in Montreal as part of a pre-budget tour. He said the federal government will begin consulting with industry soon on a new AI Compute Access Fund and an accompanying strategy to expand the sector in Canada.” So he didn’t actually announce a real investment. He announced an intent to have a budget item and to vaguely consult with…someone. Given the absolutely dire electoral prospects of the federal Liberals come the next election, this is basically meaningless.
- gxt 3y agoOver 400 billions in grants since Trudeau became PM in 2015. https://open.canada.ca/data/en/dataset/432527ab-7aac-45b5-81d6-7597107a7013 https://open.canada.ca/data/en/dataset/432527ab-7aac-45b5-81... If you're gonna do data analysis of that gigabyte CSV watch out for the 40 billions in contract adjustments. Only sum the last amendement per unique. For reference, the gross debt of the central government of the Canadian federation before him was 1000 billions. In December of 2023 it had skyrocketed to 1800 billions.
- iancmceachern 3y agoHow much of this goes straight to power. Something like 20-30% of your airline ticket goes to fuel. I wonder what the % is for AI companies
- matthewfelgate 3y agoMurr was right.