2025-05-07 12:58:59
About 88% of the value of cars made in Canada are exported to the USA, 94% of steel produced in Canada is exported to the USA, about 90% of aluminium made in Canada is exported to the USA.
2025-05-06 23:06:56
In 2022, four professors in law, computing and finance alerted Quebecers to the risks and opportunities of digital identity in Quebec. Despite our warnings, neither the political parties nor most of the media took the subject seriously. Today, we found out that a key architect of Quebec's digital identity project has been implicated in a fraud.
Your data is no joking matter. We are playing with fire.
Here is what we wrote back in 2022 (translated in English):
«««
The introduction of a digital identity for Quebecers is scheduled for 2025, which is practically just around the corner.
Several governments, including Quebec, are currently working on developing a digital identity to reduce costs, facilitate access to services, and combat fraud. Thus, digital identity is set to replace physical identity documents with a computer application.
Digital identity aims to reduce fraud and identity theft by enabling automated identity verification. If well-implemented, it could also enhance privacy protection by giving citizens greater control over how their data is used. Additionally, citizens will no longer need to remember multiple passwords. Administrative tasks such as paying property taxes, applying for or renewing a building permit, accessing medical records, or obtaining a subsidy could soon benefit from a single, secure authentication process guaranteed by the government.
Transitioning from paper-based identity documents to a digital version poses privacy risks, as every use of a digital identity could generate traceable digital records that are automatically stored. In contrast, with the current "paper format," presenting an ID to a person does not directly create long-term traces unless that person explicitly records the interaction in a system.
The ability to cross-reference data enabled by this automatic collection raises significant privacy concerns. The centralization of digitally collected citizen data allows for the creation of new data through the combination of separately collected information. Beyond improved management of certain security risks, this detailed mapping of citizens could limit their freedom or impact their privacy.
Indeed, one of the primary privacy risks associated with digital identity is the systematic tracking of citizens.
Beyond partnerships with private companies, the state could use this data to categorize citizens based on their behavior in areas such as health, mobility, or financial management. In a crisis context, such as the COVID-19 pandemic, the state might use this data to assess compliance with lockdown measures or analyze citizens’ mobility without their consent, as was done by the Public Health Agency of Canada with telephone data.
These issues become particularly acute in the context of a legal state of emergency. In such cases, the government could decide to use this data for purposes other than those initially intended, without democratic oversight.
Digital identity also raises questions about data ownership, particularly regarding who owns the data, especially data generated by cross-referencing other data. It will also be necessary to determine whether inferred data automatically belongs to the state that "produced" it.
The question of data ownership is fundamental because it determines how the data can be used. The Quebec government has expressed its intention to commercialize this data. However, it is worth asking whether digital identity and its associated data constitute a public good or a common good (a "common" in economic terms). If so, a data governance mechanism will need to be established to decide how and by whom this data can be used.
It is well-known in economics that economic actors engage in a race for innovation to achieve a monopoly and evade regulatory and economic constraints. Given that the pace of legislation often lags behind that of innovation, the question arises now about how to regulate this innovation that is digital identity.
Thus, the question is whether to anticipate and regulate potential misuses or, for the sake of operational efficiency, address issues only after they arise. Similarly, should there be a framework or model to protect data beyond existing laws, or do we accept the idea that technological innovation remains outside the law?
»»»
https://www.journaldemontreal.com/2022/06/10/lidentite-numerique-quels-enjeux-et-quels-risques
2025-05-04 04:04:14
C++20 concepts: for better error messages.
https://x.com/lemire/status/1918754843641983286
2025-05-03 01:30:27
Universities have no clue what industry needs. Industry does not know what it needs. Government would be the last to know.
We wrongly attribute the title of 'expert' to people who have no clue, no actual expertise... We confuse power and authority with expertise.
Wherever you are, you should focus on building cool stuff that you think people like. Then you test it out. If you are wrong, you course correct, if you are right, you move right ahead.
If it feels like you are doing the work of a plumber, then you are on the right track. Plumbers help people and they get paid.
Do not pay attention to what people are saying. Pay attention at what they are doing.
2025-05-01 22:34:00
The Soviet Union won the Cold War. We've been conquered. It's just that we stupidly think we've won.
I am looking at you Microsoft Teams.
2025-05-01 13:09:33
🚀 New simdutf release! 7.0.0
Enhanced base64 decoding & perf boosts!
- decode_up_to_bad_char in base64_to_binary_safe fixes JS engine issues (WebKit #290829)
- C++20 span support for cleaner syntax
- atomic_base64_to_binary_safe
- Hybrid base64 decoding (default+URL)
- Bug fixes & speed-ups
About simdutf: The simdutf library is a high-performance, open-source library designed for efficient Unicode text processing, particularly for converting between UTF-8, UTF-16, and UTF-32 and Base64 encodings. It leverages Single Instruction, Multiple Data (SIMD) instructions, such as those found in modern CPUs to parallelize and accelerate operations, achieving significant speedups compared to traditional scalar methods. The library is used by major Web browsers and JavaScript runtimes.
https://github.com/simdutf/simdutf/releases/tag/v7.0.0
credit: @PaulDreik @pshufb and Shreesh Adiga
