Lorenzo Billi

Computer Science student @ Università degli studi di Genova

I turn pizza and caffeine into software

[Lorenzo Billi @lore_billi]

@keyashqa @Pallavi_345 This is a (correct) example on why we can't *parallelize* everything. Asynchronous does not necessarily mean (and usually isn't) parallel code.

[Lorenzo Billi @lore_billi]

@_devJNS I mean, if the quality standard of most Meta's engineers is the same as the ones who shipped the latest WhatsApp webapp that consumes multiple GBs of RAM for doing literally nothing, then yes, AI will probably be better.

[dotnetliguria @dotnetliguria]

🟣 Le novità in Visual Studio 2026 – Lorenzo Billi
Nuovo Fluent design, Copilot integrato e miglioramenti di performance in VS2026.

📅 .NET Conf 2025 – 12 dicembre 2025
📍 Piazza della Vittoria 11/10, Genova
🎟️ Iscriviti 👉 https://t.co/dEz8jH1v0n

#dotnetconf

[Lorenzo Billi @lore_billi]

@DanielW_Kiwi 2) write SQL than developers. Not to mention that, by updating EF to a newer version it can improve the generated queries automatically by leveraging new DB features. Plus, of course what others have already mentioned (type safe queries, migrations...)

[Lorenzo Billi @lore_billi]

@DanielW_Kiwi 1) For EF/LINQ in particular: for the same reason we (almost) don't write Assembly by hand anymore. A compiler generates it starting from a higher level language, and it's often better at generating it than a human. EF has reached a maturity level that is often better to [...]

[Lorenzo Billi @lore_billi]

@K3vn_C @korewaznsk @Cartidise @stupidtechtakes @JailBreakMeiPod Home editions don't have Bitlocker, however disk encryption is supported anyway

[Lorenzo Billi @lore_billi]

@inkballpro @exchange12rocks @PikminGuts92 @tomwarren Nope, the host OS is based on Windows Server

[Lorenzo Billi @lore_billi]

@CodeUnloadMusic @LinusTech That's actually the space used by the page table entry, which grows proportionally with the total available RAM (since it's used to track RAM allocation), and it's not a Windows-only feature

[Lorenzo Billi @lore_billi]

@davidfowl It's not even a particularly heavy load. My 2019 XPS 15 can easily handle that, while having still quite spare power to use, and it's not even a new laptop...

[Lorenzo Billi @lore_billi]

@SandeepGaju @sundarpichai It's a Pixel

[Lorenzo Billi @lore_billi]

@jupean @tomwarren All the other major OS out there are built on even older (and often worse) technology than Windows, i.e. UNIX which is from the early '70. BTW, I can't understand why MS is creating this tool, given the fact that modern Windows (8/10/11) don't slowdown anymore with time.

[Lorenzo Billi @lore_billi]

@cate55965375 @LiFraGNUxment That P3 XL is fake

[Lorenzo Billi @lore_billi]

@peppedx @svpino I know it well, it was a shortcut for "Many libraries are written in C" / "Many libraries are written in C++"

[Lorenzo Billi @lore_billi]

@peppedx @svpino Because most if not all of the libraries used for ML/DL are written in C/C++. Even most of the Python libraries are actually wrappers of C/C++ code.

[Lorenzo Billi @lore_billi]

@TuX_Illindi @LinusTech Compared to *NIX, Windows makes large use of threads, both at the kernel level and at user level. Over 4000 threads after 20 minutes is perfectly common as a number.

[Lorenzo Billi @lore_billi]

@Suleymanzade @rasbt Grand Central Dispatch has nothing to do with GPU programming, nor with multiprocessing. It's a glorified thread pool, used for multithreading.

[Lorenzo Billi @lore_billi]

This 👇🏻👇🏻👇🏻

[Lorenzo Billi @lore_billi]

@b0rk Peco for sure:
https://t.co/zwS6KXKrUX

[Lorenzo Billi @lore_billi]

@AjaySR17 @CorentinMiq @JonyIveParody Yep, my bad, I was referring to the M1 Ultra vs RTX 3090