https://lawrencecpaulson.github.io/2025/07/02/Finish_your_degree.html

Paulson's take on billionaires and PhDs. 😁

“Relative Toposes for Artificial General Intelligence„ slides.
Part 1 and Part 2.

Hmmm... 🤔

The International Conference on Formal Structures for Computation and Deduction (FSCD) will be streamed online via Zoom for free Tuesday 15 - Friday 18 July, 2025. You can register here:
https://fscd2025.github.io/

🥦 Broccoli Programming

BroccoliController.java

@RestController
@RequestMapping("/broccoli")
@RequiredArgsConstructor
public class BroccoliController {
  private final BroccoliService broccoliService;
  private final BroccoliMapper mapper;

  @GetMapping("/get_some")
  public BroccoliDTO getSome() {
    return mapper.toDTO(broccoliService.getSome());
  }
}

BroccoliService.java

@Service
@RequiredArgsConstructor
public class BroccoliService {
  private final BroccoliRepository  broccoliRepository;

  public Broccoli getSome() {
    return broccoliRepository.getSome();
  }
}

BroccoliRepository.java

@Repository
@RequiredArgsConstructor
public BroccoliRepository {
  private final JdbcTemplate jdbcTemplate;

  public Broccoli getSome() {
    return jdbcTemplate.query(...);
  }
}

> требуется работница с сознанием компьютера
Вы находитесь здесь
> требуется компьютер с сознанием работницы
> требуется работница с сознанием компьютера

I'm still waiting for when they start employing LLMs to forecast the weather. I want to read forecasts I like, and not this crap of a weather!

Сколько кошку ни корми — шило из жопы не выдавливается...

ДМК Пресс перевели ещё одну книжку про компиляторы для начинающих:
https://dmkpress.com/catalog/computer/programming/978-5-93700-391-1/
Х. Мёссенбёк "Конструирование Компиляторов"

В девичестве "Compiler Construction": https://ssw.jku.at/CompilerBook/

Книжка про компиляторы — это хорошо, что плохо — из 8 глав 5 рассказывают про синтаксический разбор, и ещё одна — просто введение. Т.е. только две главы посвящены "мясу" компиляции: семантическому анализу и генерации кода. Для главы про оптимизации места не нашлось, к сожалению. При этом генерация кода рассматривается для стековой виртуальной машины, поэтому распределение регистров, выбор и скедулинг инструкций остались за бортом. Семантический анализ глубиной и широтой охвата похвастаться тоже не может.

Но в плюсе наличествует предисловие от Никлауса Вирта, глава про атрибутные грамматики и большое количество упражнений.

#compiler #book

If you were thinking about visiting Romania in mid-September, you can crash The Working Formal Methods Symposium as well:
https://fromsymposium.github.io/
September 17-19, Alexandru Ioan Cuza University, Iași, Romania

The end of this month will see https://conferences.au.dk/confest2025
CONFEST: August 25-30, 2025 in Aarhus, Denmark.

Microsoft open-sourced (under the MIT license) their VS Code Copilot extension:
https://github.com/microsoft/vscode-copilot-chat

There's a lot of code, and quite expectedly mostly infrastructure code. Still one can learn how they manage searches for the relevant code in the project tree, interaction with Git, analysis of the LLM responses, chats with the user and so on.

— Если проблему можно решить при помощи денег, то это не проблема, а расходы.
— Где ж взять столько денег?
— А вот это уже действительно проблема.

https://orangedatamining.com/blog/data-tells-stories-statistics-shuts-your-mouth/

Data tells stories; statistics shuts your mouth

There was an old joke (among physicists?) that there are two kinds of truths: trivial truths and fundamental truths. Trivial truth is such a statement that the opposite is false. Fundamental truth is a statement the opposite of which is also true.

The title of the post is a lot like fundamental truth...

The author (Janez Demšar) means the stories are a good and interesting thing, while statistics fanatics simply try to shut your mouth on the ground your statistics is not good enough (not rigorous enough). And Janez have very good arguments (at least examples) he presents in the post!

So I totally agree with his point. And yet, at the same time there are lots and lots of "stories" around us that pretend to be supported by data, but in reality just fantasies. And statistics really should have shut the mouths of people telling those stories. Sometimes silence is indeed golden.

https://github.com/marimo-team/marimo

An alternative to good (not really) 'ol Jupyter Notebooks.

Looks promising. First of all, they solve the biggest Jupyter problem: stateful execution and thus dependence on the order of cell evaluation. The problem is so large they trained ML models to predict the correct evaluation order. I kid you not.

The solution is pretty common and straightforward: upon loading a file, build the Dataflow Graph, then re-run all the transitive dependencies upon a cell change. This also gives you the "reactive programming for free".

Reactive updates also give you a reactive UI (almost for free). Demos look pretty nice and useful:
https://marimo.io/p/@marimo/embedding-visualizer

And while we're at it, yeah they have a Web playground, and can convert notebooks into Web pages (also slides) out-of-the-box. Moreover, they can produce dynamic Web pages without a server employing WebAssembly. I haven't tried it, but if it works well, that's immensely useful.

Among other things, they store notebooks as valid Python files, which provides integration with Git and many other tools for free. They also integrate with package managers, especially uv, and support per-project virtual environments out-of-the-box.

Additionally, they support DuckDB for running SQL queries, including on DataFrames. And implement nice UI to view, filter and summarize the results.

Overall, feels like a cool and handy, batteries included tool, alleviating a lot of Jupyter Notebooks quirks and letting you get your crap done quicker.

VS Code now has "AI settings search".

Yep, that's right, they use a freaking LLM to search in a code editor settings.

And they showcase in the release notes an example search for "increase text size" that returns font size settings. Are you sure that's a meaningful example? Do we really need an LLM to discover font size settings?

At any rate, did you consider simplifying freaking settings before implementing a LLM search through them?

I was pointed to
https://mourner.github.io/bullshit.js/

Remind everyone about
https://www.sansbullshitsans.com/

😁

https://dmkpress.com/catalog/computer/programming/978-5-93700-228/

ДМК переводит книжку про RIIR, оригинал которой недавно хвалил Виталий Брагилевский:
https://vitaly-reads.beehiiv.com/p/vitaly-reads-3-august-2025

Damn! Only now I realized the FRACTRAN was invented by John Conway himself! It emerged from his actual research.

За сутки стояния в кружке, кофе из V60 превратился в кофе растворимый Якобс Монарх.

— Духи не приняли мою жратву
— Может, жертву?
— Нет 🤢

Installing collected packages: pytz, multipledispatch, tzdata, typing-extensions, toolz, threadpoolctl, six, setuptools, pygments, pillow, packaging, numpy, mdurl, kiwisolver, fonttools, filelock, cycler, cloudpickle, cachetools, scipy, python-dateutil, markdown-it-py, logical-unification, h5py, contourpy, rich, pandas, matplotlib, h5netcdf, cons, xarray, etuples, xarray-einstats, miniKanren, pytensor, arviz, pymc

Emphasis mine. I only wanted PyMC and I've got miniKanren and logical-unification for free. Nice, but why, I wonder...