Vlad Dumbrava

Front-End Developer

[Anders Hejlsberg @ahejlsberg]

Huge milestone for our team today: TypeScript 7 is now generally available--a native port that runs 10x faster. @typescript https://t.co/3DtJJqEZAl

[Andrej Karpathy @karpathy]

LLM Knowledge Bases

Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So:

Data ingest:
I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them.

IDE:
I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides).

Q&A:
Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale.

Output:
Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base.

Linting:
I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into.

Extra tools:
I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries.

Further explorations:
As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows.

TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

[Журнал «Код» @thecode_media]

Основатель GitLab получил рецидив редкого рака кости — врачи развели руками.

Тогда он собрал 25 ТБ медицинских данных, загнал в ИИ и нашел экспериментальное лечение в Германии. Опухоль удалили. Рак отступил.

Все данные в открытом доступе 🧬

https://t.co/R2sRgQgCPc

[Cheng Lou @_chenglou]

My dear front-end developers (and anyone who’s interested in the future of interfaces):

I have crawled through depths of hell to bring you, for the foreseeable years, one of the more important foundational pieces of UI engineering (if not in implementation then certainly at least in concept):
Fast, accurate and comprehensive userland text measurement algorithm in pure TypeScript, usable for laying out entire web pages without CSS, bypassing DOM measurements and reflow

[Anthropic @AnthropicAI]

We’ve identified industrial-scale distillation attacks on our models by DeepSeek, Moonshot AI, and MiniMax.

These labs created over 24,000 fraudulent accounts and generated over 16 million exchanges with Claude, extracting its capabilities to train and improve their own models.

[NestJS @nestframework]

🚀 Big updates coming in NestJS v12:

✨ Native ESM support
⚡ Faster testing with Vitest + SWC (default for ESM projects)
🧩 Standard Schema in route decorators - use Zod, Valibot, ArkType, and more
🎨 A complete redesign of the NestJS websites

Details: https://t.co/pH3Q42RvGt

[Pavel Durov @durov]

🇲🇩 About a year ago, while I was stuck in Paris, the French intelligence services reached out to me through an intermediary, asking me to help the Moldovan government censor certain Telegram channels ahead of the presidential elections in Moldova.

After reviewing the channels flagged by French (and Moldovan) authorities, we identified a few that clearly violated our rules and removed them. The intermediary then informed me that, in exchange for this cooperation, French intelligence would “say good things” about me to the judge who had ordered my arrest in August last year.

This was unacceptable on several levels. If the agency did in fact approach the judge — it constituted an attempt to interfere in the judicial process. If it did not, and merely claimed to have done so, then it was exploiting my legal situation in France to influence political developments in Eastern Europe — a pattern we have also observed in Romania 🇷🇴

Shortly thereafter, the Telegram team received a second list of so-called “problematic” Moldovan channels. Unlike the first, nearly all of these channels were legitimate and fully compliant with our rules. Their only commonality was that they voiced political positions disliked by the French and Moldovan governments.

We refused to act on this request.

Telegram is committed to freedom of speech and will not remove content for political reasons. I will continue to expose every attempt to pressure Telegram into censoring our platform. Stay tuned. 🤝

[Homeland Security @DHSgov]

Gotta Catch ‘Em All.

[Адовый UX @ux_from_hell]

Коллекция живых форм ввода номера с адовым UX

[DG MEME 🇪🇺 @meme_ec]

When everybody is waiting for that Moldovan wine they've heard so much about.

[Пaл Лaич (Василич)💙 @king_of_botox]

ЭЙ, @MedvedevRussia, ТЫ ТАМ ЧЕГО БЛЯТЬ НАТВОРИЛ?
ИДИ ТЕПЕРЬ, ИЗВИНЯЙСЯ ПЕРЕД ПАПОЧКОЙ.
ДОЛБОЁБ СИНИЙ, БЛЯТЬ.

[Walme @walme_io]

🎉 300,000+ users in the Walme Waitlist!

This isn’t just a number — it’s the population of Zurich, one of Europe’s major cities.

Imagine: an entire Zurich has already joined the Web3 movement with Walme.
People from all over the world are signing up to be among the first to access Walme and participate in the Test TGE.

🚀 Thank you for being part of this journey — the big things are just beginning.
Not on the list yet? Join now 👉 https://t.co/oYwrIn8Eqc

[Elon Musk @elonmusk]

I regret some of my posts about President @realDonaldTrump last week. They went too far.

[Теперь в Telegram! @nowintelegram]

ФСБ теперь в Telegram!

[Konstantin Kisin @KonstantinKisin]

Yes, yes, the richest man in the world is upset about subsidies being taken away from one of his companies that he has repeatedly said should be taken away. It's definitely not politicians who repeatedly promised to stop overspending massively overspending. Definitely.

[Whale Psychiatrist ™️ @k_ovfefe2]

BREAKING: Vladimir Putin has offered to negotiate a peace deal between President Trump and Elon Musk

[Konstantin Kisin @KonstantinKisin]

Well, at least lefties will stop burning Teslas now...

[Konstantin Kisin @KonstantinKisin]

What Elon Musk is doing is genuinely heroic.

He will win no friends in politics and will be ostracised by the current administration. But he is totally right.

We (the entire Western world) simply cannot go on pretending our way of life is sustainable.

Historically, countries build up surpluses in peace time and then go into debt to fight wars. We are going into more and more debt during peace time.

Clever economists will tell you this is sustainable. It is not. And someone has to do something about it.

[Pavel Durov @durov]

🔥 This summer, Telegram users will gain access to the best AI technology on the market. @elonmusk and I have agreed to a 1-year partnership to bring xAI’s @grok to our billion+ users and integrate it across all Telegram apps 🤝

💪 This also strengthens Telegram’s financial position: we will receive $300M in cash and equity from xAI, plus 50% of the revenue from xAI subscriptions sold via Telegram 💰

Together, we win ❤️📈🏆

[Pavel Durov @durov]

Nicolas Lerner, the head of France’s foreign intelligence — who asked me to silence conservative voices on Telegram in Romania ahead of its elections (and later tried to deny it) — visited Romania just two days before the vote, official sources told journalist @tuca_ro.