Jung, Yong-Bok

researcher, rock engineering, dad of two kids

[Nein_Danke @Nein_Danke_]

“선택지가 상대적으로 많았던 식민지 엘리트에게는 시대와 구조를 앞세운다. 반면 선택지가 훨씬 적었던 위안부 여성에게서는 선택과 관계의 흔적을 집요하게 찾아낸다. 엘리트의 선택은 구조 속으로 숨기고, 피해자의 선택은 구조 밖으로 끌어낸다.”

명쾌한 분석이다.

https://t.co/6ItPHorjSJ

[公益社団法人 日本地震工学会【公式】 @JAEEGAKKAI]

地質図のビューワー。活断層の走行線と「近畿トライアングル」をデジタル地図で詳細に確認できます。
赤系は火成岩、グレーや青系は堆積岩で、地質の境界やずれも見えます。「活断層データ」をONにしてください。
https://t.co/U3C4xwxo8A

[Vaibhav Sisinty @VaibhavSisinty]

A college student just killed the biggest advantage politicians have ever had. The gap between the lie and the fact-check.

He built a Chrome extension called InTruth. It listens to any live speech, debate, or interview.

Transcribes every word. And checks every claim against real sources in real time. Before the speaker even finishes the sentence.

It works on debates. Press conferences. Town halls. Interviews. Anything with someone talking on video. The fact-check shows up on your screen while they're still talking.

Until now, fact-checking happened hours or days later. By then the clip has been shared, the headline has been written, and the damage is done. Nobody reads the correction.

InTruth closes that gap to seconds. The claim and the fact-check exist in the same moment.

A college student, a Chrome extension, and an AI model just did what entire newsrooms couldn't figure out for decades.

[金氷三 @PresidentVSKim]

만약 1993년 하나회 숙청(?) 당시 윗대가리 장군들만 몇명 전역 시키고 '친목 성격의 동호회일 뿐인데 영관급 위관급 까지 손 댈 필요야 있겠나'라고 했으면 과연 어떻게 됐을까를 생각해 보면 답이 나오재?

[욱카피 @untamarag28]

표현의 자유대로 고등학생이 그린 윤석열차 그림에 금상 줬다고 만화영상진흥원 예산 후려친 윤석열은 공산주의자란 얘기군

[南無 (Namu2Namu) @Namu2Namu]

혐오 발산
→ 문제시되면 몰랐다
→ 허접한 사과문 올림
→ 방문해서 사과하겠다고 우김

전형적인 혐오 놀이 패턴임. 광주일고도 알고 있으니까 거절함. 절대 사과 받음 안됨.

[유진 Choi @wpdlatm10]

류근

배재고등학교 야구부 학생들을 용서하자고 말씀하시는 분들 계십니다. 잘못 가르친 어른들 책임이라고…

그럼 어른들 전부를 처벌합시다.

세월호 단식 농성장 앞에서 짜장면 시켜 먹고 피자 시켜 먹던 자들 혼내주지 않은 우리 공동체의 도덕성 붕괴가 성장해서 오늘날 광주의 죽음을 떼창과 떼춤으로 모독하는 지경에까지 이르렀습니다.

용서?

전두환, 이명박, 박근혜를 용서해서 우리 사회가 1센티미터라도 좋아졌나요? 참담하고 암담합니다. 어른답게 분노하고 혼을 내야지요. 촉법 제도에 대한 사회적 숙의 필요합니다. 일베짓이 놀이가 되는 세상이라니요!

광주 5.18에 총맞아 죽은 중3 아들 주검 앞에서 울부짖는 엄마 사진에 대고서, 아이고~ 홍어 택배 왔네! 이렇게 놀고 있는 쓰레기들에게 아무런 불이익이 없다는 게 믿어집니까? 저런 놈들이 또 윤석열 같은 악마를 키우는 겁니다. 혼내야 해요.

[다행엄마 @zzikkuki]

배재고 사건은 교육 문제 아님. 한국사회 질적 하락을 막느냐 못 막느냐의 문제임. 애들한테 너무 하는 거 아니냐는 말에 동의하지 않음.

[먹물이 @mugmulree]

배재고 덕에 고교야구판 큰 피바람 불겠네

배재고는 일단 광주일고한테만 그런게 아니라 광주 동성고, 광주 진흥고한테도 똑같은 짓을 한게 드러나버렸고 배재고 뿐만 아니라 충암고 1학년들도 같은 짓을 했다는 증언이 속속 올라오는 중 모르고 했다.

놀이일 뿐이라고 하기엔 광주지역 고등학교만 골라서 한걸 봐서 얘네들 이거 뭔지 알고 한거임

아무래도 10대들이 인스타, 유튜브 쇼츠를 통해 구구들이 뿌리는 일베문화를 쉽게 접하고 있는게

드디어 큰 문제를 만든 듯 10대들이 일베문화를 하나의 놀이로 여긴다는게 실제로 드러나버렸음

[Millie Marconi @MillieMarconnni]

A British novelist whose father worked in a bank explained the entire 2008 financial crisis in a 272-page book that a 12-year-old can follow, and the part that should scare you is how simple the fraud was once he stripped the jargon off it.

His name is John Lanchester.

He is a novelist by trade. He writes for the London Review of Books and The New Yorker. He grew up in Hong Kong because his father worked at a bank there, which means he absorbed the language of finance the way most kids absorb the language of cartoons. By the time the 2008 crisis hit, he was one of the very few writers on earth who could speak both finance and English fluently.

He sat down and wrote a book called I.O.U.: Why Everyone Owes Everyone and No One Can Pay.

In the UK it was published under the title Whoops! It came out in 2010. 272 pages. The reviews from financial journalists were almost embarrassed in their praise, because Lanchester had done in plain English what most of them had failed to do in years of professional coverage. He had actually explained what happened.

Here is the version a middle schooler can understand.

Step one. A bank gives a home loan to a person who clearly cannot pay it back. No real job. No savings. No way to make the monthly payments. The industry had a nickname for these loans. They called them NINJA loans. No Income. No Job. No Assets. The bank knew the person could not pay. The bank did not care.

Why did the bank not care?

Because the bank was not going to keep the loan.

Step two. The bank takes thousands of these loans, puts them in one giant pile, and sells the entire pile to a Wall Street firm. The Wall Street firm now owns the right to collect all those monthly payments. The original bank has been paid. Whether the homeowners actually pay or not is now someone else's problem.

Step three. The Wall Street firm takes the giant pile of loans and chops it into slices. They call the slices tranches. The top slice is supposed to be the safest because it gets paid first. The bottom slice is supposed to be the riskiest because it gets paid last. They give each slice a credit rating. The top slice gets stamped AAA, which is the highest rating you can get on earth. It is the rating used for the bonds of the United States government.

A pile of loans to people with no income and no job has just produced an investment product rated as safe as the United States government.

Step four is where the trick gets worse.

The bottom slices, the ones full of the riskiest loans, are hard to sell. Nobody wants junk. So the Wall Street firm takes the bottom slices from a hundred different piles, mixes them together into a new pile, and chops that new pile into new slices. And the top slice of this new pile, made entirely from the junk of the old piles, also gets stamped AAA.

This new product has a name. It is called a Collateralized Debt Obligation, or CDO.

A 12-year-old can see what is wrong here. You cannot take a hundred bags of garbage, dump them into one big bag, and then declare that the top of the big bag is now gold. The garbage is still garbage. The math does not change it.

But the people who ran the ratings agencies were not 12-year-olds. They were adults with finance degrees. And they stamped this stuff AAA over and over again because the Wall Street firms were paying them to do the stamping. The ratings agencies had become a paid service. The people they were rating were also the people writing their checks.

Step five. The Wall Street firms sell these AAA-rated CDOs to pension funds, foreign governments, insurance companies, and retirement accounts all over the world. Everyone buys them because the rating says they are safe.

Step six is the final trick.

Someone invents a product called a Credit Default Swap. It is insurance against a CDO going bad. The clever part, the part that Lanchester explains so well, is that you do not have to actually own the CDO to buy insurance on it. You can buy insurance on a CDO that belongs to someone else. You can buy insurance on a CDO ten times over. A hundred times over. There is no limit.

This is the equivalent of taking out fire insurance on your neighbor's house. And then on a thousand strangers' houses. And then betting against the entire neighborhood.

The smart people in the industry realized the CDOs were garbage. So they started buying credit default swaps against them. Not to protect themselves. To bet on the collapse.

Step seven was the part nobody planned for.

Housing prices stopped going up. The NINJA borrowers, who had been refinancing every year by using the rising value of their houses as collateral, suddenly could not refinance. They started missing payments. The bottom tranches of the CDOs started failing. Then the middle. Then the AAA tranches that had been declared safer than government bonds turned out to be made of the same garbage as the rest.

Then the insurance bills came due.

The biggest insurer in the world, AIG, had written hundreds of billions of dollars in credit default swaps on these CDOs. They did not have the money to pay. If AIG collapsed, every major bank on earth that had bought insurance from them would also collapse, because their "safe" CDOs would suddenly have no insurance. The American taxpayer ended up writing AIG a check for 182 billion dollars to keep the entire global financial system from falling apart.

The total damage of the crisis, by every honest estimate, runs into the trillions of dollars. Millions of people lost their homes. Millions lost their jobs. Almost nobody on Wall Street went to jail.

Here is the part Lanchester drives home in the final chapters, and the part you should not forget.

The fraud was not hidden in the math. The fraud was hidden in the complexity. Each individual step looked legitimate. Each individual product had a name, a rating, a paper trail. The reason nobody stopped it was that no single regulator, no single executive, no single auditor was forced to look at the entire chain from end to end. The garbage at the start and the AAA stamp at the end were separated by so many layers of paperwork that everyone could honestly say their specific layer looked fine.

Lanchester's central argument is the one that should haunt every person reading this in 2026.

Complexity is the disguise. Anytime a financial product is so complicated that the person selling it cannot explain it in 60 seconds, the complexity itself is the warning sign. The complexity is not protecting the customer. The complexity is hiding the fact that the customer is the product.

The 2008 crisis was not a failure of intelligence. It was a failure of plain language.

Almost every product on the market today that is being sold to retail investors, from leveraged ETFs to certain crypto derivatives to private credit funds, has the same structural pattern. Layered. Hard to explain. Stamped with an authoritative rating. And the people selling it can never quite tell you what is inside.

Which brings me to the question.

The next 2008 is already being built somewhere in the financial system right now.

Which industry, or which product, do you think is hiding the same trick under the same kind of complexity right now in 2026, and what makes you suspicious of it?

[How To Prompt @HowToPrompt__]

Stanford + Meta just dropped the paper that flips everything about AI agents.

It's called "Code as Agent Harness."

Right now, we treat large language models as text generators. When they need to solve a complex problem, they rely on a "chain of thought."

But natural language is slippery. It's vague. It loses context. When an agent hallucinates in English, it just keeps talking.

So they introduced a framework that changes the entire architecture of autonomy: "Code as Agent Harness."

They stopped asking the AI to reason in words, and forced it to reason in code.

Code isn't just the final output anymore. It is the memory. It is the environment. It is the boundary.

Instead of writing a paragraph about how to solve a problem, the agent writes a script, executes it, and reads the output.

Tests become its senses. Execution logs become its memory. Sandboxes become its physics.

If an agent makes a mistake in English, it apologizes and hallucinates again.

If an agent makes a mistake in code, the compiler throws an error. The trace tells it exactly what broke. The system forces it to fix it.

This is where prompt engineering dies, and systems engineering takes over.

The paper proves that reliability doesn't come from a smarter base model. It comes from the "harness" wrapped around it:

- The model proposes.
- The harness executes.
- The environment returns feedback.
- The verifier checks.

[Xudong Han @Xudong07452910]

这可能是人类写给 AI 看的最后一篇论文了。

最近刷到Stanford、CMU、Michigan 等 37 位作者联名的论文：《The Last Human-Written Paper》。

核心观点很狠：沿用几百年的论文，在 AI 时代可能已经过时了。

作者点出了两个被我们忽视已久的“隐形税”：

一个是叙事税。为了讲一个漂亮故事，我们把失败实验、死路、被推翻的假设都删掉了。AI 读到的是“通关攻略”，却看不到真正有价值的“踩坑记录”。

另一个是工程税。论文里的实现细节通常足够说服审稿人，但不够让 Agent 直接复现。很多关键 tricks 还藏在作者脑子、代码注释和 Slack 记录里。

所以作者提出 ARA，直接把论文改造成 Agent 能读取和执行的“研究包”：不只告诉你结论，还把怎么想到的、代码怎么跑、证据链在哪、哪些路走不通都打包进去。

我觉得这篇最有意思的地方是，它不是在讨论 AI 怎么帮人写论文，而是在问：

当 AI 也变成论文读者和执行者时，论文还应该长成今天这样吗？

未来科研输出的核心，可能不再是“写得多像一篇 paper”，而是能不能被 AI 理解、复现、追踪和继续扩展。

人类写论文写了几百年，接下来可能要开始写给 Agent 执行的研究包了。

https://t.co/2wwkWV9Ox4

[Zara Zhang @zarazhangrui]

If you've adopted AI at your company but haven't seen any tangible results, read this 1990 article: "The Dynamo and the Computer" by Paul David.

When electricity first arrived, factories that "adopted" it barely got faster. They just swapped the steam engine for an electric one and ran everything else exactly as before: same machine layout, same workflow, same management. Electricity in, no real gains out.

The most common mistake with any new technology is to drop it into the old organization and then declare the transformation done.

The real leap came decades later, when each machine got its own small motor. Suddenly machines no longer had to be lined up around one central drive shaft. They could be rearranged around the actual flow of work.

The productivity gains didn't come from electricity. They came from REDESIGNING THE ENTIRE FACTORY around it.

AI is the same. Bolting it onto your existing process gets you a faster steam engine. The payoff comes when you redesign the work itself.

(link to paper in comments)

[Spencer Baggins @bigaiguy]

A French engineer who lives quietly in Paris has spent 30 years writing software that the entire internet now runs on without knowing his name.

He wrote the code that streams every YouTube video, every Netflix show, every TikTok clip. He wrote the code that runs the virtual servers underneath AWS, Google Cloud, and Microsoft Azure. He calculated more digits of pi than anyone in history. He has no Twitter. He has no marketing. He just keeps shipping.

His name is Fabrice Bellard.

Here is the story, because almost nobody outside the systems programming world knows what one man has built.

Fabrice was born in 1972 in Grenoble, France. He studied at École Polytechnique, the top French engineering school. He never went to Silicon Valley. He never built a startup empire. He just wrote code.

In 2000 he started a project called FFmpeg, an open-source multimedia framework for encoding, decoding, and streaming video. He was 28. The project did one thing nobody else had done well. It handled every video and audio format that existed, in one library, on every operating system. He led it himself for years.

Today FFmpeg is the invisible engine of the internet. YouTube uses it. Netflix uses it. VLC uses it. Chrome and Firefox use parts of it. Every Android phone, every iPhone, every smart TV, every video editing tool you have ever touched runs FFmpeg somewhere underneath. If you have watched a video on a screen in the last 20 years, Fabrice's code processed it.

He was not done.

In 2003 he started QEMU, a machine emulator and virtualizer. He wrote it solo until version 0.7.1 in 2005. QEMU lets you run any operating system on any other operating system. It became the foundation of modern virtualization. KVM, the Linux kernel hypervisor, runs on top of QEMU. Every major cloud provider, AWS, Google Cloud, Microsoft Azure, IBM Cloud, runs virtual machines on infrastructure built around it. The Quick Emulator is the most cited piece of cloud infrastructure code on Earth.

He kept going.

In 2001 he won the International Obfuscated C Code Contest with a small C compiler that grew into TCC, the Tiny C Compiler. TCC can compile and boot a Linux kernel from source in under 15 seconds. In 2004 he calculated the most digits of pi ever computed at the time, using a personal desktop computer and an algorithm he derived himself called Bellard's formula. In 2011 he wrote a complete PC emulator in pure JavaScript that runs Linux in your browser, a project called JSLinux that engineers still cannot believe is real.

In 2019 he released QuickJS, a small but complete JavaScript engine that fits where V8 cannot. In 2021 he released NNCP, a neural network based lossless data compressor that immediately took the lead on the Large Text Compression Benchmark.

Then he turned his attention to large language models. He built TextSynth Server, a web server with a REST API for running LLMs locally. He released ts_zip and ts_sms, compression utilities that use language models to compress text and short messages at ratios traditional algorithms cannot reach. He released TSAC, a very low bitrate audio compression system. In December 2025 he released Micro QuickJS, a new JavaScript engine for microcontrollers, separate from QuickJS, designed for environments with almost no memory.

Fabrice co-founded a telecom company called Amarisoft in 2012, where he serves as CTO. Amarisoft builds 4G and 5G base station software used by carriers and labs around the world. He has been running it for over a decade while continuing to ship personal projects from his own home page at bellard dot org

He has no Twitter. He has no Instagram. He gives almost no interviews. His personal website is a flat list of projects with no styling, no fonts, no marketing copy. Just titles and links.

A quiet French engineer who never moved to Silicon Valley wrote the code that quietly runs the internet.

He is still shipping.

[Notion @NotionHQ]

So @_Brian_Zhang built a full-on simulation of The Office using Notion Custom Agents in < 48 hours.

He took 1st place, received 10k+ in credits, then sat down with us to explain how it all worked. We chat about:

> Proximity based reasoning for agents
> Managing long term context windows
> Keeping characters “in character” with reflection loops
> The push/pull between plan mode and reactive decision-making
> What surprised him most

Michael would be proud.

[Movez @0xMovez]

Claude Code creator:

"I don’t prompt Claude anymore. What I mostly use now is loops. I create loops - they do the rest of my job."

In 24 minutes, Boris reveals his real daily Claude Code setup:

Claude + loops + routines + dynamic workflows

Worth more than a $500 vibe-coding course

[Ejaaz @cryptopunk7213]

claude mythos just broke Apple's $2 billion defense system. it did so by discovering a completely different attack vector to break in

only took it 5 days costing ~$35K of mythos api time (the same exploit class costs $5-10M on grey market)

the researchers that commandeered the exploit produced a 55-page report that was delivered to Apple HQ in-person (hoping they release it after patching).

most shocking part for me is apple's MIE worked as intended. mythos just discovered a new way to side-step it entirely by poisoning the data the M5 chip ingested.

at this point i think we have to accept that mythos walks the walk.

As the anthropic red-team explicitly confirmed this week - this is NOT a compute resource issue. its national defense.

[Lancy @LNCV34]

메모.

"모든 민원을 동등한 무게로 처리하는 시스템은 결국 가장 시끄러운 자에게 보상을 주는 시스템."

https://t.co/tfjjIFHd1z

[한겨레 @hanitweet]

“만약 당신의 삶에서 딱 하나만 바꿀 수 있는 “마법의 카드”가 주어진다면 무엇을 바꾸겠는가? 학벌, 직장, 아파트…. 다양한 상상이 가능하겠지만, 지금 대한민국에서 가장 잔인할 만큼 현실적인 답은 하나다. 부모를 바꾸는 것. 이 비극적인 상상은 노력이라는 신화가 파산한 우리 시대의 자화상이다.”
https://t.co/ap7nurzsPr

[Crazy Clips @crazyclips_]

Switzerland really set the bar for weather presenters