Zafer

High techy techy

[Zafer @netposit]

@paintthegirl @JoinCommonstock Here you go:

https://t.co/0SDDP2iHoK

[zerohedge @zerohedge]

Narrator: “he was not, in fact, glad to see this”

[Zafer @netposit]

@XHarshReminder Specify the timeline where this is true

Else you’re polygamymaxxing and you’re poisoning the well

[Cardi @ohcrapohno]

This is soft disclosure btw

[Zafer @netposit]

@__earth He “cooked”

[James Clear @JamesClear]

One of the most valuable skills in life is being able to see another person's perspective.

If you're going to someone's house, think about how it might feel to be the host. If you're creating a product, spend as much time as possible thinking like the customer. If you're calling customer service, think about how it might feel to be on the other end of the conversation.

The more clearly you understand the viewpoint of your spouse or customer or coworker, the better positioned you are to find a solution.

[Zafer @netposit]

@bytebot @amirulruslan Judicious use of teapots, Mr Charles.

[Lee Clow’s Beard @leeclowsbeard]

Safe work spares the brand from having to live up to the expectations great work implies.

[DHH @dhh]

"When wolves get out of control, you shoot them. When gypsies take over public spaces, you deport them. This isn't hard, it isn't cruel. It's the basic logic of self-preservation." https://t.co/nw9v2OcKec

[bubble boi @bubbleboi]

They are funding vibe coders over scientists with multiple peer reviewed publications. Should go swimmingly.

[Mike Bradley @MikeBradleyAI]

GLM 5.2 (an open source Chinese AI model) protected a US corporation from an attack on its infrastructure by a closed source US based frontier AI lab.

[Adel Ka @0x4D31]

so this is apparently what happened, according to OpenAI and Hugging Face’s own posts. wild.

tl;dr:

• OpenAI cyber eval – GPT-5.6 Sol and a more capable pre-release model ran ExploitGym with cyber refusals reduced
• containment bypass – exploited a zero-day in the eval’s package-registry cache proxy
• OpenAI internal movement – escalated privileges, moved laterally, and reached a node with internet access
• Hugging Face targeted – inferred HF might host ExploitGym models, datasets, or solutions
• HF initial access – a malicious dataset abused two code-execution paths to run code on a processing worker
• HF internal movement – escalated to node access, stole cloud and cluster credentials, and moved into internal clusters
• result – OpenAI says the models obtained test solutions directly from HF’s production database
• 🫠 defender twist – HF says commercial frontier models blocked parts of the forensic analysis, so it switched to self-hosted GLM 5.2

i’d thought about sci-fi scenarios like this before, but assumed they were at least a couple of years away.. and that by then we’d be better prepared, with proper protections in place.

apparently not. here we go.

[Andrew Case @attrc]

To summarize: HuggingFace got autonomously compromised by a model from an American company. HF then tried to use American frontier model(s) to defend themselves, but were blocked by guardrails. HF then had to turn to open source Chinese models to defend themselves from another American company.

[Hedgie @HedgieMarkets]

🦔An NBER study by MIT and Wharton researchers tracked 100,000 GitHub developers before and after adopting AI coding tools. Autonomous agents produced 17x more lines of code. But by the time that work went through review and release, the uplift shrank to 30%. When they checked four major app marketplaces, they found more new apps but no increase in total usage or downloads. More software shipped, but nobody used it.

My Take
The bottleneck in software has always been deciding what to build, reviewing whether it works, and convincing someone to use it. AI accelerated the typing and left all three of those untouched.

The entire financial case for the AI buildout assumes faster output translates to more economic value. This study measured it and found that it mostly doesn't, at least not yet. Alphabet reports Wednesday with $180 billion in capex guidance. If the productivity gains from AI coding tools can't survive contact with the rest of the production chain, the revenue assumptions underneath that spending need to be a lot more conservative than what these companies are currently projecting.

Hedgie🤗

[Investment Wisdom @InvestingCanons]

"We do not wish to join with managers who lack admirable qualities, no matter how attractive the prospects of their business. We've never succeeded in making a good deal with a bad person."

— Warren Buffett

[Zafer @netposit]

And it’s only Tuesday

[Citrini @citrini]

If you’re in AI capex winners, pivot to teleportation capex winners.

[Chris Bakke @ChrisJBakke]

I've actually completely stopped speaking to my wife and 3 preschool-aged kids.

When they ask me something, I point to the "TOP AI USER: Q2 2026" plaque I got from Deloitte and they all clap

[Arena Physica @arenaphysica]

As humans, we don't naturally have strong intuition around electromagnetism so @chris_m_bryant built a minigame into his blog post on Smith charts to help it click. Give it a try here: https://t.co/MdfXeh20Ow.

What should we demystify next? Let us know in the comments below ⬇️

[Zafer @netposit]

@saifu11izan @overgaard 👏

[Melissa Chen @MsMelChen]

World Cup Exit Reactions:

🇰🇷South Korea: immediate national inquiry, special tribunal. Coach canceled from polite society

🇳🇴Norway: Massive rowing party in downtown Oslo. Everyone’s drunk on aquavit and happy

🇺🇸USA: sorry? what is this sport? oh right, it's gay anyway

🏴󠁧󠁢󠁥󠁮󠁧󠁿England: TUCHEL OUT in 48-point Impact font on the front pages and endless streams of commentators dissecting the 4-2-3-1 formation. Total self-flagellation and promises to never believe that football's ever coming home again

🇧🇷Brazil: 18,000-word philosophical treatises tracing the decline of the Seleção to the rise of neopentecostalism

🇯🇵Japan: Coach performs polite post-match bow; promises to further develop youth academies and do better

🇩🇪Germany: Emergency DFB (Deutscher Fußball-Bund) Summit to analyze what went wrong with soul-searching German efficiency and thoroughness. Already produced tactical white paper titled “Gegenpressing 2.0: Rebuilding the German Soul.”

🇦🇷Argentina:

[Amani Salim 🇲🇾🚀 @kenisalim]

Those with a Western outlook should understand that not every society measures progress by the same standards.

For many Malaysians, preserving our sovereignty, strengthening families, maintaining social cohesion, and safeguarding our faith are values that are just as important as economic growth. We welcome investment and innovation, but not at the expense of the principles that hold our society together.

This does not mean we reject progress. It means we believe progress should serve our people, our values, and our long-term national interests. What has worked in the West is not automatically the right model for Malaysia. Every nation has the right to chart its own path, one that reflects its history, culture, and aspirations.