Lightnews — Scholar-powered news

Felipe O. Carvalho @felipe.rs · 14h

This is a common thing and the reason behind avoiding pointers to register-sized values across function boundaries. An add10() that returns the value added by 10 is better than the one that mutates through the pointer. Totally fine if inlined though.

1

Felipe O. Carvalho @felipe.rs · 3d

If we forbid the inlining, the caller is forced to put i on the stack

str r0, [sp, #4]

to pass the ptr to add10

add r0, sp, #4 // put the ptr relative to the stack ptr on r0 before call

then the value is loaded from the stack into r1 before the printf call

ldr r1, [sp, #4]

$__attribute__((noinline)) int add10(int *num) { return *num += 10; }$

3

Felipe O. Carvalho @felipe.rs · 3d

add10(int *) versus add10 inlined. When you take a "pointer" from a local variable and "pass" it to a function, but the compiler knows the pointer is not a pointer but the register that contains the local variable.

$#include <cstdio> int add10(int *num) { return *num += 10; } int external_int(); int main(int argc, char *argv[]) { int i = external_int(); add10(&i); printf("%d", i); return 0; }$

2 3

Felipe O. Carvalho @felipe.rs · 4d

It’s only a 20-ish% increase from June in Brazil, so I’m securing mine before they have to renew their inventory and increase it even more.

380 BRL for Kingston Furybeast 16GB in June (~70 USD)

1

Felipe O. Carvalho @felipe.rs · 5d

Oh, second ed is out already?

2

Felipe O. Carvalho @felipe.rs · 5d

When an LLM finally confesses it just “speculates based on general patterns” 😅

Marcus Hutchins @malwaretech.com · 5d

bruh

2

Felipe O. Carvalho @felipe.rs · 5d

Ordered 32GB of DDR5 today. 💵

It’s been a long while since I bought RAM that is not the one soldered in a Macbook’s motherboard. Feels good and affordable.

1 7

Felipe O. Carvalho @felipe.rs · 7d

De longe

Felipe O. Carvalho @felipe.rs · 9d

The number of numbers in red circles I have throughout all software I use at work is maddening. Zoom is a big offender here trying to get me to read chats from all meetings.

Seanan McGuire @seananmcguire.bsky.social · 10d

I legitimately hate how every website seems to think that what I want more than anything else in this world is yet another inbox to monitor and take responsibility for.

1 7

Felipe O. Carvalho @felipe.rs · 16d

It's annoying in C++ to query an std::unordered_map<std::string, T> when what you have is a std::string_view (pointer, "borrowed") instead of the std::string (allocated on the heap, "owned").

1 2

Felipe O. Carvalho @felipe.rs · 16d

Rust solved the problem of querying a map structure without necessarily allocating a key value. The key type can have a borrowed version that is equivalent to the owned version of the key.

1 4

Felipe O. Carvalho @felipe.rs · 16d

“wE lAunChed our SaaS a mONth ago and alReady mADE it to 1M ARR”

Friendly reminder that multiplying revenue by 12 is not ARR

1 2

Felipe O. Carvalho @felipe.rs · 17d

Add to Playlist gets worse every month or so

1 2

Felipe O. Carvalho @felipe.rs · 22d

Elon is bragging about X not being affected by the outage today and, unfortunately, his audience can’t appreciate that all the decisions that lead to this were taken by the team he laid-off when Twitter was acquired.

1 2

Felipe O. Carvalho @felipe.rs · 23d

Wondering if this is the kind of thing that @welltypedwit.ch will say is not good in practice.

There is a nice online demo/playground: se-tuebingen.github.io/oopsla-2025-...

The Simple Essence of Overloading: Playground

se-tuebingen.github.io

Felipe O. Carvalho @felipe.rs · 23d

The idea seems very applicable to programming systems that offer a high-level API that is supposed to be implemented in terms of implementations that are specialized for different types for perf reasons (e.g. ML frameworks, dataframe libraries)

dl.acm.org/doi/10.1145/...

The Simple Essence of Overloading: Making Ad-Hoc Polymorphism More Algebraic with Flow-Based Variational Type-Checking | Proceedings of the ACM on Programming Languages

Type-directed overload resolution allows programmers to reuse the same name, offloading disambiguation to the type checker. Since many programming languages implement overload resolution by performing...

dl.acm.org

1 3

Felipe O. Carvalho @felipe.rs · 29d

AI coding tools make us exploit all the shortcuts you can to ship more quickly at the cost of eventually making it unsustainable.

2 3

Felipe O. Carvalho @felipe.rs · Oct 6

100% this!

Felipe O. Carvalho @felipe.rs · Oct 5

Scary how completely unconcerned they are with re-election. Final years of the term will be devoted to diminishing the electoral system itself.

1

Felipe O. Carvalho @felipe.rs · Oct 4

I had these before moving to the Microsoft Sculpt. Sculpt is not as durable as this one, but this keyboard is way too big and takes over the whole desk. Happy with a mechanical split these days.

1 1

Felipe O. Carvalho @felipe.rs · Oct 4

In a 100 years, people will be talking about these companies like we talk about IBM during WWII.

2

Felipe O. Carvalho @felipe.rs · Oct 4

It’s really sad to see people still doing this at this point. It’s as if they get a high from figuring out the hypocrisy.

Felipe O. Carvalho @felipe.rs · Sep 29

I love how he said he doesn’t feel bad for being somewhat disconnected from the current developments of AI because he is paying much more attention to what thinkers in other areas are talking about cognition and learning.

2

Felipe O. Carvalho @felipe.rs · Sep 27

e.g. compression and crypto algorithms are very sophisticated, but if you hide that in the HTTP stack, people working on a web app don't have to understand their details when implementing applications.

1

Felipe O. Carvalho @felipe.rs · Sep 27

I'm not a fan of the approach, but I've seen very pedestrian codebases survive and succeed a lot more than the codebases that had sophisticated constructs that got perverted by changes. Note that sophistication is not forbidden. It's requiring everyone to understand it to work on the project.

1