2025-04-04 04:04:56
Large language models pass as human beings more often than human beings themselves.
In other words, you can already replace human beings by AI and nobody will notice.
How sure are you that I wrote this post?
2025-04-03 04:15:10
My main laptop is a 16GB macbook (not pro) and though I can run out of RAM and have to reboot… it is still fine… mostly.
I should point out that a lot of my code runs on remote servers (cloud).
2025-04-03 00:45:09
In Mississipi, they had talked about cutting the income tax whenever the state income exceeded the state expenses by 85%. Someone wrote 0.85% instead of 85% and it is now the law.
2025-04-02 20:19:50
Recent disks (SSDs) by Kingston and others can read and write data at ~15 GB/s. And we are on the verge of doubling these bandwidths.
2025-04-01 03:22:46
I was asked whether I expect Canada to soon become competitive in the AI race—particularly in the context of some Canadians advocating for infrastructure independent of the USA.
My answer is concise: no.
Canada had the potential to be a global leader in AI. It invested significant funding and boasts some of the world’s top engineers and researchers. Yet, despite these advantages, it has fallen short of that promise.
A key reason lies in Canada’s cultural priorities, which do not emphasize technical innovation. This stems, in part, from the "resource curse." The country’s economy leans heavily on exporting raw materials—oil, minerals, lumber, and other commodities—rather than fostering cutting-edge industries like AI.
In practice, launching an ambitious high-tech venture in Canada comes with significant challenges. Entrepreneurs face regulatory hurdles, limited venture capital, and a broader ecosystem that doesn’t support scaling innovative technologies. These headwinds stifle the kind of bold, independent AI development one needs.
Government funding is not necessarily helpful in this context. To get government grants, compliance with various high level societal objectives is of paramount importance. Market success if often quite secondary. This tends to generate a subsidized entrepreneurial culture where entrepreneurs are selected for their ability to secure grants, rather than to actually build services or products.
In turn, these subsidized ventures can capture and occupy promising young people... sometimes leaving them older and demoralized.
It is definitively possible to grow a culture of technical innovation in Canada, but it is not as simple as creating a government program.
You can succeed in Canada. There are intriguing cases, such as @Shopify or the (now dead) BlackBerry. And a few others. Canada played a role in the IT revolution. But no matter how you square it, it played the role of a vassal of the USA.
I do not expect an anti-USA sentiment, no matter how strong, to have a sizeable effect on this issue.
2025-03-29 10:11:17
Supporting obsolete processors is costly.
If you are publishing an accounting software and some of your customers run Windows 7… it is fine to build your software for legacy systems.
However, video games and graphics software should safely assume a decent processor… i.e., AMD Zen 2 or better.
If you can pay for a modern video game, you should be able to pay for decent hardware.
Supporting legacy systems is complicated and slows the engineering.
