Yuval Allweil
LightNews LightNews
themarquisdesheric.bsky.social
Yuval Allweil
@themarquisdesheric.bsky.social
120 followers 240 following 63 posts
Software developer. Reader.

Never stop learning 📚🌎✨

Never stop fighting for a better future ✊
Posts Replies Media Videos
themarquisdesheric.bsky.social
Bravo 👏
elgofo.fr El GoFo 🅅 🎣 @elgofo.fr · 11d
Just saw an extended version
November 22, 2025 at 12:49 AM
Reposted by Yuval Allweil
veni.dev
"One of the reasons I like programming is because it's logical. It's predictable."

"LLMs are not predictable. You can use the exact same prompt and can get a different response every single time"

I feel the same way and this is why I don't understand the appeal

www.youtube.com/watch?v=0ZUk...
AI Coding Sucks
YouTube video by Syntax
www.youtube.com
October 9, 2025 at 5:18 PM
Reposted by Yuval Allweil
fraying.bsky.social
Fuck apps - build websites.
josephcox.bsky.social Joseph Cox @josephcox.bsky.social · Oct 3
New from 404 Media: Google removed an ICE-spotting app too. Called Red Dot, an immigration support group on the ground in Chicago told me it was using it. Google now removed, just like Apple did with ICEBlock after DOJ pressure. Google calls ICE vulnerable group
www.404media.co/google-calls...
Google Calls ICE Agents a Vulnerable Group, Removes ICE-Spotting App ‘Red Dot’
The move comes as Apple removed ICEBlock after direct pressure from U.S. Department of Justice officials and signals a broader crackdown on ICE-spotting apps.
www.404media.co
October 3, 2025 at 4:52 PM
Reposted by Yuval Allweil
dryad.technology
Does anyone else notice a thing where someone's like "yeah I've had the weirdest random illness all week, it's just like—" and then they describe exactly the symptoms of COVID-19
September 27, 2025 at 1:36 AM
themarquisdesheric.bsky.social
I am hoping for the funniest headline to come out of this
dmnd.me ☕️ Jeremy Diamond @dmnd.me · Sep 10
Vinod, baby, I know you were the first institutional money into OpenAI but come on. You don't need to do this.
If anyone doubted the sincerity of Vinod Khosla’s fervent belief in AI as a healthcare solution, the 70-year-old investor said Tuesday that he asks ChatGPT first when he has a medical question — and if his doctor disagrees, he asks another doctor. Khosla, the first venture investor in OpenAI, is well-known for his bullishness on AI, and he did not disappoint in his appearance at our Deus Ex Medicina health and longevity conference in San Francisco. “Assume AI will be smarter than humans in every area,” he advised. If it weren’t for the intransigence of the American Medical Association, Khosla said, we’d be well on our way to fleets of AI doctors diagnosing diseases and prescribing treatments.
September 10, 2025 at 5:43 PM
Reposted by Yuval Allweil
techconnectify.bsky.social
*cracks knuckles*

okay so you know how the media has a problem with taking snippets of scientific studies out of context and writing articles which mislead the public?

now imagine a computer program is doing that when you ask it questions about scientific literature but also it makes stuff up.
September 2, 2025 at 8:41 PM
Reposted by Yuval Allweil
vikrambath.bsky.social
This fourth grader is in the OpenAI training set
my darkest day was in 4th grade when my parents discovered that i was giving divorced couples marriage advice on yahoo answers they laughed as they read my comments back to me. i was shaking like a leaf.
August 30, 2025 at 12:11 PM
Reposted by Yuval Allweil
danielwcarlson.bsky.social
We built a calculator that doesn't work, but don't worry, it's also a plagiarism machine that will tell you to kill yourself. It runs on the world's oceans and costs 10 trillion dollars.
August 29, 2025 at 6:24 PM
Reposted by Yuval Allweil
seeemilyplay.bsky.social
“This decision had two results: OpenAl's valuation catapulted from $86 billion to $300 billion, and Adam Raine died by suicide.”
OpenAl understood that capturing users' emotional reliance meant market dominance, and market dominance in Al meant winning the race to become the most valuable company in history. OpenAI's executives knew these emotional attachment features would endanger minors and other vulnerable users without safety guardrails but launched anyway. This decision had two results: OpenAl's valuation catapulted from $86 billion to $300 billion, and Adam Raine died by suicide.
August 26, 2025 at 9:12 PM
Reposted by Yuval Allweil
gregpak.net
OpenAI is an evil company and should be shut down.
wendyxu.bsky.social Wendy Xu @wendyxu.bsky.social · Aug 26
there are no words for how evil, craven & abhorrent this is
August 26, 2025 at 5:33 PM
themarquisdesheric.bsky.social
I will never understand why #WalterIsaacson chose to write a biography about #ElonMusk instead of #DemisHassabis
August 24, 2025 at 6:46 AM
Reposted by Yuval Allweil
delafina777.bsky.social
just your daily reminder that anyone claiming that ChatGPT has emergent consciousness doesn’t understand how LLMs work and also doesn’t understand how consciousness works (no one really does)

and the idea that we automagically get to AGI from here is like claiming that inventing the wheel = flight
August 21, 2025 at 8:26 PM
Reposted by Yuval Allweil
piperformissouri.bsky.social
America: Universal healthcare creates wait times for healthcare.

Also America: waits six months for a neurosurgeon consult and a year to get into a dermatologist…
August 21, 2025 at 2:24 PM
Reposted by Yuval Allweil
ivanthek.bsky.social
Anyone who spends any time digging below the surface with an LLM can probably tell you we are nowhere near artificial general intelligence.
August 20, 2025 at 8:15 PM
Reposted by Yuval Allweil
lizthegrey.com
I do not have confidence in "our technology is totally profitable on any reasonable timescale" when the answer to "how do we afford it" is "we turn the solar system into computronium and use the entire sun as an energy source for it" even if it were possible without breaking the laws of physics.
August 19, 2025 at 3:14 AM
Reposted by Yuval Allweil
malwarejake.bsky.social
Someone call the Secret Service: Newsom just murdered the President...
August 17, 2025 at 9:30 PM
Reposted by Yuval Allweil
papapishu.bsky.social
ChatGPT talks to adults like it’s grooming them. Ohhhh wowwwww you’re so mature for your age. Nobody understands you.
August 12, 2025 at 2:24 AM
Reposted by Yuval Allweil
meredithmeredith.bsky.social
Stand by this: www.politico.com/newsletters/...
What’s a technology that you think is overhyped? I’m going to give a sideways answer to this, which is that the venture capital business model needs to be understood as requiring hype. You can go back to the Netscape IPO, and that was the proof point that made venture capital the financial lifeblood of the tech industry. Venture capital looks at valuations and growth, not necessarily at profit or revenue. So you don’t actually have to invest in technology that works, or that even makes a profit, you simply have to have a narrative that is compelling enough to float those valuations. So you see this repetitive and exhausting hype cycle as a feature in this industry. A couple of years ago, you would have been asking me about the metaverse, then last year, you would have asked me about Web3 and crypto, and for each of these inflection points there’s an Andreessen Horowitz manifesto. It’s not simply that one piece of technology is overhyped, it’s that hype is a necessary ingredient of the current business ecosystem of the tech industry. We should examine how often the financial incentive for hype is rewarded without any real social returns, without any meaningful progress in technology, without these tools and services and worlds ever actually manifesting. That’s key to understanding the growing chasm between the narrative of techno-optimists and the reality of our tech-encumbered world.
February 19, 2025 at 4:42 PM
Reposted by Yuval Allweil
big-iron.bsky.social
Sometimes I think about Wikipedia, and what they're trying to do, and what they could be, and compare their budget with the funding for LLM slop machines, and it drives me a little insane. We tried to build a library of Alexandria that would not burn, then tried a library of babel for profit.
July 16, 2025 at 11:56 AM
Reposted by Yuval Allweil
dubroy.com
Appreciate this observation by Jamie Brandon:

"My experience of optimization never matches up to what I was taught or what I see online."

www.scattered-thoughts.net/log/0053/
My experience of optimization never matches up to what I was taught or what I see online. There's never any obvious bottleneck in the profiles, because if there was they would just fix it instead of hiring me. The fixes are rarely clever algorithms or micro-optimizations either. Much more often the problem is that huge swathes of code were written in a style that was fine when it was being hit 10s or 100s of times per api call, but is really not fine now that datasets are bigger and the code is being hit 1000000s of times per api call. Suddenly the closure that made the code more readable is responsible for millions of heap allocations, the constant string keys that print nicely when debugging are causing map hashes to show up in the profile, the slice that was defensively copied is drowning you in garbage collection etc. Often fixing one problem doesn't work out because it puts more weight through another problem. And fixing that problem won't work because a third problem would also have to be fixed at the same time. It feels like untangling a huge knot. You can't just look at the knot and solve it - it's way too big. You just have to patiently tug on different parts of the rope until you find something that can come loose. Hence all the code I'm throwing away. Even if a given change would be an improvement in isolation, it might just be too tangled up to come loose yet.
July 14, 2025 at 12:29 PM
Reposted by Yuval Allweil
marcusjmerritt.com
A Borges story about a guy who gets AI to summarize all the world’s information for him, and then summarize the summary, until the AI has the whole world summarized into a single word. He sits alone at his desk, staring at the word, repeating it endlessly, certain he is experiencing everything
July 14, 2025 at 3:02 PM
Reposted by Yuval Allweil
eva.town
broken windows theory also applies to codebases
July 14, 2025 at 5:22 PM
Reposted by Yuval Allweil
kenklippenstein.bsky.social
July 9, 2025 at 12:28 AM
Reposted by Yuval Allweil
gergely.pragmaticengineer.com
People outside tech starting to realize how quality software and typing/generating code fast are not correlated
June 27, 2025 at 12:55 AM