Aderyn

🌊
Building with 🤖

[Motivaciones Fútbol @MotivacionesF]

¡ESTA IMAGEN NOS ROMPIÓ EL CORAZÓN! 💔🇮🇷

Ramin Rezaeian marcó el gol del empate, fue elegido como el mejor futbolista del partido… pero eso no alcanzó.

Tras el pitazo final, no pudo contener las lágrimas. Irán estuvo a segundos de conseguir la clasificación, pero el gol de la victoria en el último minuto fue anulado.

La imagen de Ramin recibiendo el premio con los ojos llenos de lágrimas lo dice todo. Refleja el dolor de un equipo que no solo luchó dentro de la cancha, sino también fuera de ella.

Mientras otras selecciones viven el Mundial con normalidad, Irán ha tenido que entrenar en México y viajar el mismo día de cada partido a Estados Unidos. Un contexto que ha complicado enormemente su participación.

Esto fue mucho más que un partido de fútbol.

TIENEN TODO NUESTRO RESPETO, IRÁN. 🇮🇷❤️

[Open Source Intel @Osint613]

Russian air defenses over Moscow in action. This footage is absolutely crazy.

[Aderyn @_9th_Wave]

@ggreenwald Yea, Israel needs to wake up and realise that they have become pariahs and it's so undeniable now that even the US government has to acknowledge these facts. Enough is enough with their bloodlust

[Glenn Greenwald @ggreenwald]

I just made this my ring-tone.

It could all lead to nothing, or worse (a joint US/Israel resumption of the war). But there hasn't been a week where American leaders have spoken so sternly, clearly, truthfully and decisively about Israel since . . . well, perhaps ever.

[Ryan Grim @ryangrim]

Some common-sense things Trump said today that were nevertheless shocking:

Everybody else has ballistic missiles, why can’t Iran?

The frozen money belonged to Iran, and the U.S. needed to release it otherwise nobody would trust the dollar system. Why should we steal their money?

Neighboring countries use nuclear power for electricity, why can’t Iran?

Israel shouldn’t knock down entire buildings in Beirut every time a drone or two crashes harmlessly in the desert.

Trump would’ve become Herbert Hoover in a few weeks once oil ran out and the market crashed. He had no choice but to end the war.

[Keanu Nazari 📕 @KeanuNazari]

I got thousands of rare Cultural Revolution-era photos from a historical archive here in China.

I believe some of these have never been posted online, or only exist behind extremely expensive licenses.

So here they are, for free. Open the thread for more.

[David Walpiri @DWalpiri]

Argentine Navy aircraft recorded hundreds of Chinese fishing vessels creeping through it's EEZ.

the sheer number of lights across the water like an invading locust plague. These predators are going to strip the entire oceans bare for CPC's endless appetite

[Aderyn @_9th_Wave]

@gbrumfiel @latimes @nabihbulos Apparently footage of this happening - the guy flees his car and then you hear an explosion
https://t.co/IxpjqN3oRy

[Lord Bebo @MyLordBebo]

🇮🇱🇱🇧 Man from southern Lebanon received a phone call from the Israeli army informing him he was a target & ordering him to exit his vehicle

Otherwise, they said he would be killed along with his family

Man exited the car, and ran into a nearby field

Then, he was subsequently targeted by a drone strike ...

[Andrej Karpathy @karpathy]

Judging by my tl there is a growing gap in understanding of AI capability.

The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code.

But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along.

So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions.

TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.

[Generative History @HistoryGPT]

Opus 4.7 is very good at historical handwriting recognition. On our modified metrics (ignore capitalization and punctuation differences) it scores 3.44% word errors and 1.76% character errors. This is around a 40% improvement from Opus 4.6 which is remarkable for an iterative release.

[GREG ISENBERG @gregisenberg]

this video is the CLEAREST explanation of how claude skills + AI agents work and how to use them

most people set up an AI agent and wonder why it keeps disappointing them.

the context window is everything

context is what the model assembles before it takes any action. think of it like everything the agent needs to read before it does anything. the quality of what goes in determines the quality of what comes out. the models are genuinely really good right now. claude and gpt are exceptional. the variable is almost always the context you give them.

1. agent.md files are mostly unnecessary

every single line you put in an agent.md file gets added to every single conversation you have with your agent. a 1000 line file is around 7000 tokens burning on every run. the model already knows to use react. it can read your codebase. save the agent.md for proprietary information specific to your company that the model genuinely cannot know on its own.

2. skills are the actual unlock

a skill.md file works differently. what loads into context is only the name and description, around 50 tokens. the full instructions only appear when the agent recognizes it needs that skill. so instead of 7000 tokens on every run you have 50. and the agent stays sharp because the context window stays lean. the closer you get to filling the context window the worse the agent performs, same way you perform worse when someone dumps 10 things on you at once.

3. here is how to actually build a skill the right way

most people identify a workflow and immediately try to write the skill. what you want to do instead is run the workflow by hand with the agent first. walk it through every single step. tell it what to check, what good looks like, what bad looks like. correct it in real time. once you have had a full successful run from start to finish, tell the agent to review everything it just did and write the skill itself. it writes a better skill than you will because it has the full context of what actually worked in practice not in theory.

4. recursively building skills is how you go from frustrated to reliable

when the skill breaks, and it will break, ask the agent exactly why it failed. it will tell you specifically what went wrong. fix it together in that same conversation. then tell it to update the skill file so that failure mode never happens again. ross mike did this five times with his youtube report generator. it now pulls from eight different data sources and runs flawlessly every single time without him touching it.

5. sub agents are something you earn not something you set up on day one

start with one agent. build one workflow. turn it into one skill. once that works add another. ross mike has five sub agents now covering marketing, business, personal and more. it took months to get there and every single one exists because a workflow proved it deserved to exist. the people who set up 15 sub agents on day one and wonder why nothing works skipped all the steps that make the thing actually run.

6. your workflow is the thing the model cannot get anywhere else

the model has been trained on everything. it knows more than you about most things. what it does not have is your specific process, your taste, your way of doing things. that is what skills capture. that is what makes your agent actually useful versus a generic one. downloading someone else's skill means downloading their context onto your setup and it will not work the way you want it to because it was never built around how you work.

this is the clearest explanation of how agents actually work i have heard. @rasmic runs this stuff every single day and the results show it.

full episode is now live on @startupideaspod where you get your pods

people charge for this sorta stuff

i give away the sauce for free

i just want you to win

watch

[OSINTdefender @sentdefender]

Video footage showing an alleged Iranian ballistic missile impact during the start of Operation Epic Fury, overnight on February 28, against Naval Support Activity Bahrain, the home of the U.S. Navy’s Fifth Fleet Headquarters, located within the Kingdom of Bahrain’s capital Manama on Awal Island in the Persian Gulf.

[OSINTdefender @sentdefender]

Earlier today, the B1 Bridge on the Karaj Northern Bypass in western Tehran, Iran was targeted in a U.S. and/or Israeli strike, amidst an uptempo in joint U.S.-Israeli strike operations, announced by U.S. President Donald J. Trump last night.

[Open Source Intel @Osint613]

CRAZY FOOTAGE 🔴

Footage reportedly shows an Iranian ballistic missile detonating just meters from U.S. troops as they take cover at an American base on March 30.

[Amirhessam.y @amirhessam_y]

بافت مسکونی میدان رسالت 💔
خدای من...

[شراره @Dehshiri4Law]

در بدترین کابوسهایمان هم این روزها نبود.

[Fazel Hawramy @FazelHawramy]

Unbelievable video

filmed by a resident from Tehran shows the reality of life under Israeli/US strikes

[‎نابخشودنی @ardalanmousavi]

برای مخالفت با جنگ مجبور نیستیم از کثافت‌کاری‌های جمهوری اسلامی حمایت کنیم.
برای حمایت از جنگ هم مجبور نیستیم چشممون رو بر قربانیان غیرنظامی ‌و بیگناه جنگ ببندیم.
#DigitalBlackOutIran #IranWar‌

[مِدیچی🎒 @cathrinemedici]

این صنایع فولاد با بدبختی و سال‌ها تلاش، زیر بار تحریم ساخته شدن.
هیچ کشورکی در خاورمیانه این حجم از دانش و تخصص رو نداره و در دنیا معدود کشورهایی همچین تاسیساتی دارند.
هر خشتی که از این‌ها پایین میریزه دل و قلب من هم باهاش میلرزه.
اینا سرمایه‌های کشورن.

[Image Trust @Image_Trust]

از امروز عکس، عکاس های داخل ایران رو که دارن این جنگ رو پوشش میدن به اشتراک میذارم لطفا به اشتراک بذارید تا فاجعه ی انسانی دیده بشه. عکاس وحید بیات:شنبه ۸فروردین ۱۴۰۵ بمباران تهران
28/03/2026-Vahid Bayat photos-Tehran