tech, eng, builder, short-term pessimist, long-term optimist
🌍
Joined March 2018
- Tweets6.4K
- Following2.1K
- Followers1.1K
- Likes33.9K
Dear governments, if you want us all to have emergency kits and a few days’ worth of supplies so badly, just hand them out already.
The test has no fingers pressing on the screen while closing it. That’s not how people always close their phones. If Apple tested it internally the same way, without accounting for this, that’s a huge oversight and could explain why it’s so fragile in real use.
I’ve had the @Apple AirPods Pro 3 for 2-3mo now, and I can say they’re performing worse than gen 2 all around, even if the fit test passes. Both sound quality and ANC are significantly worse.
If you want to buy them, make sure the store accepts a return if you are not satisfied. The stores here do not accept returns (they found a loophole related to hygiene so no headphones are accepted for returns), so I’ll have to go through warranty a few times until they decide to just give me my money back. And I simply don’t have the spare time and energy for this.
“The instrument (the telescreen, it was called) could be dimmed, but there was no way of shutting it off completely.”
— George Orwell, Nineteen Eighty-Four (1949), Part One, Chapter 1.
Samsung has basically built the infamous telescreens from Orwell’s 1984, except now the surveillance is automated.
To add insult to injury, you not only pay for the telescreen, but also for the electricity it consumes processing your data locally before uploading it.
The conversational behavior feels very natural, unlike most audio/video content you find online.
Makes you wonder where they found enough recordings of genuinely spontaneous conversations to train on.
We believe voice is the most natural way to interact with intelligence, so we created Dessa, the most authentic full-duplex speech model ever built. Talking with it feels like talking with a person…full of emotion, nuance, pauses, and a natural cadence.
block.xyz/dessa
iOS 27: Apple Health now lets you choose how far back an app can read: your full history, or just the last 30 days.
Workout Exporter for AI supports this. It tells you when the limit affects an export, and notes it in the file, so your AI won't mistake it for "no workouts".
This reminds me of how North Korea survived the zombie apocalypse in World War Z.
Spoiler alert: they pulled everyone’s teeth within 24 hours. “No teeth, no bite. No great spread.”
During the spread of Blaster worm in the early 2000s, another worm, Welchia, used the same vulnerability to enter vulnerable systems, remove Blaster, _and_ install Microsoft’s security patch to close the vulnerability.
Claudiu retweeted
If you are wondering whether your website is affected, we have packaged a Codex skill to help check.
Point your Codex at this repo to guide you through it: github.com/openai/php2shell-…
OpenAI models helped researchers—operating as part of our Patch the Planet initiative—discover critical security flaws in PHP, an open-source programming language used to run an estimated 70% of websites.
Over the past two months, we led a coalition of maintainers, researchers, and security providers to prepare fixes and protections ahead of disclosure. This duration was necessary because the fix was quite intricate and required a major change to the PHP engine. Every site using PHP can now safely upgrade their version to stay secure, in addition to applying mitigations from various security vendors.
Due to the severity of these flaws and PHP’s widespread use, we consider this bug to be one of the most impactful vulnerabilities in recent memory. It also marks a new phase, where AI can find security bugs with internet-scale impact.
This is an example of the Defender’s Window in action: to our knowledge, frontier AI was able to find this vulnerability ahead of attackers, giving defenders time to prepare.
openai.com/index/coordinated…
Readers added context they thought people might want to know
‪OpenAI may claim a role in the coordinated response and patching process. But the initial discovery and exploit development, however, occurred several months earlier by an independent researcher.‬
‪x.com/lcfr_hax/statu…‬
Claudiu retweeted
If we want to continue progress in science, we must be prepared to acknowledge that the correct description of nature will eventually simply be too complicated for humans to understand. I do not think it is wise to make human understanding a requirement. AI is the way forward.
Claudiu retweeted
We now have a good definition of the singularity:
"When AI systems speed up the pace of overall scientific and technical progress in a way that is hard for humans to comprehend."
According to Jack Clark, it's already here. I would find it hard to disagree.
ANTHROPIC CO-FOUNDER TELLS THE IMF THE SINGULARITY HAS ALREADY BEGUN
"I believe that this year, we've gained convincing evidence that it has now begun"
"a point where our models must be discarded and a new reality rules"
"the singularity is here"
at least this time Clark shares his definition of what "singularity" means, and even by this definition I'm not sure we are there yet. AI is clearly speeding up progress dramatically in some areas, but it is still uneven, with plenty of bottlenecks preventing broad, self-sustaining acceleration
even the recent math progress still looks more like a massive speed-up of technical work than completely new theories or brilliant conceptual breakthroughs. so if you ask me, the singularity is not here yet, but the current pace may already be enough to create massive social and technological turbulence
For a long time the industry was biased toward younger engineers. But now that knowledge is almost instantly available to everyone, experience, judgment, and good taste matter more than ever. And these take time to develop.
software engineering is so back.
> people are reading books again.
> every line of code is being read carefully.
> fundamentals are cool again.
> architecture and system design are back in focus.
> nobody is fighting over frameworks anymore.
> understanding why something works matters again.
> code reviews matter more than lines of code shipped.
> debugging is becoming a superpower again.
> technical judgment matters more than typing speed.
> engineering is about thinking again, not just coding.
If you, as an individual, open tens of PRs per day and no human ever reads them, maybe it’s time to switch to committing straight to main.
Claudiu retweeted
My friend Leo traveled to Tokyo carrying his late grandfather’s 1970s Seiko watch.
It had stopped working years ago. Leo didn't speak a word of Japanese, so he relied entirely on Google Translate.
In a quiet back alley, he walked into a tiny watch repair shop with just an elderly man sitting behind a cluttered desk.
Leo held up his phone and typed:
"It stopped ticking years ago. Can you fix it?"
The old watchmaker examined the watch, paused, and put on his magnifying loupe. He didn't say a single word.
He opened the caseback, took a deep breath, and worked in complete silence for 40 minutes.
Suddenly, a faint, steady ticking filled the room.
Leo teared up. He typed on his phone and showed the screen:
"How much do I owe you?"
The old man gently shook his head.
He slowly typed something on his own phone and turned the screen toward Leo:
"No charge. Look inside the caseback. That tiny engraved mark was mine. I assembled this watch 53 years ago as a young apprentice. Thank you for bringing my boy home."