Justin Trudeau announces $2.4B for AI investments in Canada
cbc.ca
cbc.ca
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.
What else should the Canadian government do? How do other more successful countries use such money to jumpstart some industrial sector?
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/202...
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.
Instead ArriveCan part deux would be a better money pit.
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.
>> 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.
Markets are a useful tool. But like LLMs they can make mistakes.
Also, why not put those brilliant young minds to work here instead, rather than give their best years to the competition?
That's brain drain, my friend
https://www.ft.com/content/0cdf78f9-e2cc-48ff-ba65-027f1cf83...
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.
Also, the government is pushing some kinda harsh AI regulations at the same time. It's kinda counter intuitive.
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.
Permission to fail is a big component of that in the US
highly addictive and privacy invading powerhouses.. investors like that?
the versions on state life support have that too.
Some more discussion: https://news.ycombinator.com/item?id=39962194
Something like 20-30% of your airline ticket goes to fuel.
I wonder what the % is for AI companies
Not sure how much it costs to make chatGPT4 quality, isnt it just including human reinforcement? Maybe that will get expensive.