What is Hackers' Pub?

Hackers' Pub is a place for software engineers to share their knowledge and experience with each other. It's also an ActivityPub-enabled social network, so you can follow your favorite hackers in the fediverse and get their latest posts in your feed.

0
1
0
1
0
0

Yesterday I did Sunday Brunch at Salty's overlooking the Columbia River and even though food was $90/person it was an insanely fun experience to spend a couple hours inside a timewarp, feeling like I was at a high end casino buffet in 1987. This review nails the vibes youtube.com/watch?v=E3m0u9H4tCA

0

Unlocking Python's Cores:Energy Implications of Removing the GIL

Link: arxiv.org/abs/2603.04782
Discussion: news.ycombinator.com/item?id=4

arXiv logo

Unlocking Python's Cores: Hardware Usage and Energy Implications of Removing the GIL

Python's Global Interpreter Lock prevents execution on more than one CPU core at the same time, even when multiple threads are used. However, starting with Python 3.13 an experimental build allows disabling the GIL. While prior work has examined speedup implications of this disabling, the effects on energy consumption and hardware utilization have received less attention. This study measures execution time, CPU utilization, memory usage, and energy consumption using four workload categories: NumPy-based, sequential kernels, threaded numerical workloads, and threaded object workloads, comparing GIL and free-threaded builds of Python 3.14.2. The results highlight a trade-off. For parallelizable workloads operating on independent data, the free-threaded build reduces execution time by up to 4 times, with a proportional reduction in energy consumption, and effective multi-core utilization, at the cost of an increase in memory usage. In contrast, sequential workloads do not benefit from removing the GIL and instead show a 13-43% increase in energy consumption. Similarly, workloads where threads frequently access and modify the same objects show reduced improvements or even degradation due to lock contention. Across all workloads, energy consumption is proportional to execution time, indicating that disabling the GIL does not significantly affect power consumption, even when CPU utilization increases. When it comes to memory, the no-GIL build shows a general increase, more visible in virtual memory than in physical memory. This increase is primarily attributed to per-object locking, additional thread-safety mechanisms in the runtime, and the adoption of a new memory allocator. These findings suggest that Python's no-GIL build is not a universal improvement. Developers should evaluate whether their workload can effectively benefit from parallel execution before adoption.

arxiv.org · arXiv.org

0
0

It's been one of the greatest honors of my life to serve as Jay's COO. She is a rare human who sees a more hopeful future and leads with integrity, grace, and wisdom. I can't wait to see how she continues to change the world as Bluesky's CIO.

RE: https://bsky.app/profile/did:plc:oky5czdrnfjpqslsw2a5iclo/post/3mgnkkbkw2c2v

0
0
0
0
0
0

ついでにXcodeを更新したみたい(Tahoeにはしたくないよう…

$ gcc --version
Apple clang version 17.0.0 (clang-1700.6.4.2)
Target: arm64-apple-darwin24.6.0
Thread model: posix
InstalledDir: /Library/Developer/CommandLineTools/usr/bin

0
0
0
0
0
0
0
0

I love and use AI heavily, but recognize and respect the human touch points. One of the things I'm proud of is that our release notes for Ghostty are still lovingly created by hand by our maintainer team. We spent over 16 hours preparing them. I hope you liked them. ❤️

0
0
0
0
0
7
0
0
0
0

A desalination plant supplying drinking water to 30 villages.

A flour and starch factory.

Oil depots blanketing Tehran in thick black smoke and contaminated rain.

65 schools. 13 hospitals.

Centuries-old heritage sites.

These are the latest targets in the illegal US-Israeli assault on Iran. These are not military bases, they are the civilian infrastructure that keeps ordinary people alive. More than 1,300 people have already been killed since the attacks began, and nearly 6,700 civilian sites have reportedly been struck.

This war has nothing to do with democracy — or even nuclear weapons. Canada must have no part in it and Prime Minister Carney must use every ounce of his influence to put a stop to it.

0
0
0
0
0
0
0

LeanTutor: A formally-verified AI tutor for mathematical proofs. ~ Manooshree Patel, Rayna Bhattacharyya, Thomas Lu, Arnav Mehta, Niels Voss, Narges Norouzi, Gireeja Ranade. arxiv.org/abs/2506.08321v1

arXiv logo

LeanTutor: A Formally-Verified AI Tutor for Mathematical Proofs

We present LeanTutor, a Large Language Model (LLM)-based tutoring system for math proofs. LeanTutor interacts with the student in natural language, formally verifies student-written math proofs in Lean, generates correct next steps, and provides the appropriate instructional guidance. LeanTutor is composed of three modules: (i) an autoformalizer/proof-checker, (ii) a next-step generator, and (iii) a natural language feedback generator. The first module faithfully autoformalizes student proofs into Lean and verifies proof accuracy via successful code compilation. If the proof has an error, the incorrect step is identified. The next-step generator module outputs a valid next Lean tactic for incorrect proofs via LLM-based candidate generation and proof search. The feedback generator module leverages Lean data to produce a pedagogically-motivated natural language hint for the student user. To evaluate our system, we introduce PeanoBench, a human-written dataset derived from the Natural Numbers Game, consisting of 371 Peano Arithmetic proofs, where each natural language proof step is paired with the corresponding logically equivalent tactic in Lean. The Autoformalizer correctly formalizes 57% of tactics in correct proofs and accurately identifies the incorrect step in 30% of incorrect proofs. In generating natural language hints for erroneous proofs, LeanTutor outperforms a simple baseline on accuracy and relevance metrics.

arxiv.org · arXiv.org

0

You can attach images, video files or audio files to Mastodon posts:

🖼️ Images can be PNG, JPG, HEIF, WEBP, AVIF or GIF (including animated GIFs)

📺 Video clips can be MP4, M4V, MOV or WebM files

🔈 Audio clips can be MP3, OGG, WAV, FLAC, OPUS, AAC or 3GP files

The maximum file size for each image is usually 16 MB, for video or audio it's usually 99 MB. There is no restriction on running time, just file size.

More info on attachments at fedi.tips/how-do-i-post-images

0
0
0
0

You can attach images, video files or audio files to Mastodon posts:

🖼️ Images can be PNG, JPG, HEIF, WEBP, AVIF or GIF (including animated GIFs)

📺 Video clips can be MP4, M4V, MOV or WebM files

🔈 Audio clips can be MP3, OGG, WAV, FLAC, OPUS, AAC or 3GP files

The maximum file size for each image is usually 16 MB, for video or audio it's usually 99 MB. There is no restriction on running time, just file size.

More info on attachments at fedi.tips/how-do-i-post-images

0
0
0

Annoyingly I seem to have tennis elbow despite not playing tennis. Maybe not initially caused by mouse use but I'm pretty sure it's exacerbated by mouse use.

Anyway, I'm training myself to use the mouse on the left. I thought I'd look at some papers on the best approach. This was an interesting find: apparently training on the non-dominant side can improve performance on the dominant side. So everyone should be doing this, injury or not :)

nature.com/articles/s41598-021

0
0
0
0

I'm a software professional and I have spent the last 40 years writing code almost daily. To me, the advent of LLM coding tools means that the job that I have been doing so far is gone. I am no longer going to write code for a living. Instead, my job as a software professional will be to guide LLMs to write the code for me.

I did not ask for this, and I have not decided whether I enjoy my new activity more than I enjoyed my previous duties. But is that relevant? In a certain sense, I am a 1/

0
0
0

M1 MacBook Airちゃんから練習 :saba:

71f9763e68 (upstream/main, upstream/HEAD) Fix erratic scroll-to-right on delete & redraft in Advanced UI (#38116)
1d46558e8d Ingestion of remote collection items (#38106)
2c6d072175 Extract `ErrorResponses` from application controller (#38105)
e235c446c9 Fix broken option defaults in `tootctl email-domain-blocks` (#38107)
4211b1b34c Update dependency faker to v3.6.1 (#38074)
73d97153d8 Update `connection_pool` to version 3.0.2 (#38103)
fcc3fac8a8 Fix `Style/ReduceToHash` cop (#38088)
3a796544e3 Allow removing yourself from a collection (#38096)
5838fce8f4 New Crowdin Translations (automated) (#38100)
3b1705a406 Add `TermsOfService#usable_effective_date` to DRY up view logic (#38095)
d7a1bead51 Update dependency aws-sdk-s3 to v1.215.0 (#38067)
eabbda3148 Fix `Style/SelectByKind` cop (#38084)
5090036edf Update dependency hiredis-client to v0.27.0 (#38092)
6a6a97f314 Update dependency json-schema to v6.2.0 (#38093)
2ad0b32dd5 Add basic ES-enabled index/service coverage (#38097)

0