Leonel Nunez (@leonelnc) header
Leonel Nunez (@leonelnc) profile picture

DevelRebel, Sysadmin

Joined August 2009
Photo by Leonel Nunez (@leonelnc)Photo by Leonel Nunez (@leonelnc)Photo by Leonel Nunez (@leonelnc)Photo by Leonel Nunez (@leonelnc)Photo by Leonel Nunez (@leonelnc)Photo by Leonel Nunez (@leonelnc)Photo by Leonel Nunez (@leonelnc)Photo by Leonel Nunez (@leonelnc)Photo by Leonel Nunez (@leonelnc)Photo by Leonel Nunez (@leonelnc)
Leonel Nunez retweeted
A Primetime performance 👏 @usbank | #FTTB
178
1,123
239
5,786
106,620
Elon Musk: "Instead of teaching to the problem, currently people teach to the tool. It would be like having a course on screwdrivers or a course on wrenches, but not understanding why."
184
436
64
2,855
969,082
More AI-generated vulnerability reports mean more triage for maintainers, and AI-generated patches can introduce regressions that only experienced reviewers catch. @IBM Chief Client Innovation Officer of Enterprise Security, Jamie Thomas @tjamieibm shared how @openssf security tools and shared standards, alongside Akrites’ incident response support, can help projects prioritize real vulnerabilities and get fixes upstream. Her call to action: “It’s our responsibility as organizations to start contributing these remediations upstream in open source.” #OSSummit
3
7
3,371
here are my favourites from the International Landscape Photographer of the Year 2025
18
119
4
1,829
200,180
Leonel Nunez retweeted
Replying to @leostera
Go is much uglier.
1
1
33
Leonel Nunez retweeted
I think about this post every time I see someone on here bragging about a perf improvement w agents. “The people who blindly trust agent output are in the former camp. They're sheeple, overdrinking from a fountain of mediocrity.” I love AI too but you need to think!
I've got an agent in a loop optimizing a renderer with the goal to minimize frame times (and tests to measure). It got times down from 88ms to 2ms and allocations down from ~150K to 500. Sounds good, right? Wrong. This is exactly why agent psychosis is a big fucking problem. As an experiment, I rewrote the Ghostty core render state in Go, with access to identically laid out data structures as Ghostty and the exact same validation tests. I made a purposely naive renderer (simple, correct, but slow). 88ms per frame with 150,000 allocations (horrendous, lol)! I then kickstarted a Ralph loop to bring the frame times down. I told it it can't modify input data structures or the public API or tests (they're correct), but it can do anything else it wants. It got to work. It has worked for about 4 hours. I've spent around $350 on this experiment so far. The results? 88ms => 1.5ms 150K allocs => ~500 allocs Incredible right? Nope. My hand-written renderer I ported has frame times (same benchmark) of ~20us (0.020ms) and 0 allocations in the update path. This is the problem with psychosis and lacking systems understanding. If you don't understand the system, you're going to accept that this is an incredible result. If you understand the system, you'll see better solutions immediately and can do roughly 75x better on throughput. The people who blindly trust agent output are in the former camp. They're sheeple, overdrinking from a fountain of mediocrity. Standard disclaimer: I use AI all the time. I like AI. The point I'm making is to not blindly accept results. Think. Analyze. Learn.
28
29
1
534
84,004
Leonel Nunez retweeted
I've got an agent in a loop optimizing a renderer with the goal to minimize frame times (and tests to measure). It got times down from 88ms to 2ms and allocations down from ~150K to 500. Sounds good, right? Wrong. This is exactly why agent psychosis is a big fucking problem. As an experiment, I rewrote the Ghostty core render state in Go, with access to identically laid out data structures as Ghostty and the exact same validation tests. I made a purposely naive renderer (simple, correct, but slow). 88ms per frame with 150,000 allocations (horrendous, lol)! I then kickstarted a Ralph loop to bring the frame times down. I told it it can't modify input data structures or the public API or tests (they're correct), but it can do anything else it wants. It got to work. It has worked for about 4 hours. I've spent around $350 on this experiment so far. The results? 88ms => 1.5ms 150K allocs => ~500 allocs Incredible right? Nope. My hand-written renderer I ported has frame times (same benchmark) of ~20us (0.020ms) and 0 allocations in the update path. This is the problem with psychosis and lacking systems understanding. If you don't understand the system, you're going to accept that this is an incredible result. If you understand the system, you'll see better solutions immediately and can do roughly 75x better on throughput. The people who blindly trust agent output are in the former camp. They're sheeple, overdrinking from a fountain of mediocrity. Standard disclaimer: I use AI all the time. I like AI. The point I'm making is to not blindly accept results. Think. Analyze. Learn.
334
1,183
265
10,617
1,000,558
Leonel Nunez retweeted
OhMyDebn 4.8.0 now installs on LCOS and Devuan Excalibur! You can now have Omarchy themes and tiling options on these distros (and the other distros listed below) Also new: • Brave Origin default browser on fresh installs • Codex CLI (optional) • Remote Desktop (optional) • ohmydebn-doctor • custom keybindings @LundukeJournal @DevuanOrg
2
6
2
20
1,717
Leonel Nunez retweeted
It's never too late to discover your courage.
108
396
27
5,281
128,932
Leonel Nunez retweeted
1
7
30
2,122
Leonel Nunez retweeted
Gunicorn 26.1.0 is released 🎉 Thanks again to the $GUNICORN community for the support. It helped me dedicate more time to Gunicorn these last days and get this release out with a good number of fixes and improvements. Now back to building :) github.com/benoitc/gunicorn/…
9
9
2
40
2,019
Leonel Nunez retweeted
Splashdown confirmed! Congratulations to the entire SpaceX team on the 13th flight test of Starship!
1,260
4,440
804
34,271
2,390,899
Leonel Nunez retweeted
Background on the Starship story
1,326
2,664
240
18,583
2,707,325
Leonel Nunez retweeted
Qualifying is always hair-raising in Monaco! 🤯💨 Heading onboard with Kimi for his @pirellisport pole position lap 🙌 #F1 #MonacoGP
115
1,298
157
9,140
1,282,506
We'll soon announce the universal distribution package that will be implemented in #Vendefoul. It needs to work on any init system. Say no to the imposition of systemd! Stay tuned! distrowatch.com/table.php?di…
7
13
1
53
1,348
@vendefoulwolf, a distribution that unites the experience of @debian (grandfather), the strength of @DevuanOrg (father), and the youthful energy (child). This is how the lessons of self-control, freedom, and privacy are forged. Creating and learning while looking to the future.
4
12
44
970
Leonel Nunez retweeted
😆😆
🤖 Made with AI
836
7,204
573
47,799
2,878,364
Leonel Nunez retweeted
“We prefer to write songs about things that people don’t do very often. Like death." 🎥Iron Maiden: Burning Ambition - tickets now available, in cinemas May 7th Head to ironmaidenfilm.com for more information.
35
211
9
1,472
63,902
this is unironically the right way to build a startup
87
189
16
4,567
148,083