Romlib 🎄

103.5K posts

Romlib 🎄 banner
Romlib 🎄

Romlib 🎄

@romlib_

Taxes are theft and that's why they're good! 🇷🇴

A place the universe forgot Katılım Mayıs 2021
1.9K Takip Edilen7.3K Takipçiler
Celeste
Celeste@celestepoasts·
cancelled and got a fairphone instead...
English
6
0
10
226
Celeste
Celeste@celestepoasts·
it's so crazy how america fumbled this guy. He so obviously really, really cares.
Celeste tweet media
English
3
1
12
205
Romlib 🎄
Romlib 🎄@romlib_·
@1nterchangeable @ChadNotChud I was curious what made delaniac tweet this since I feel like he has been using them extensively a lot too, so maybe he knows something I dont. It doesn't seem implausible to me that LLMs stagnate, but my priors just dont go that way right now.
English
1
0
0
27
Romlib 🎄
Romlib 🎄@romlib_·
@1nterchangeable @ChadNotChud My entire job is writing code with LLMs now, it wasn't a year ago. I've been coding for over a decade, and I've seen plenty of other very experienced people use LLMs over the past year and notice this improvement directly. So yeah, the massive jump feels very real to me.
English
2
0
1
36
delaniac 🌹🌱
delaniac 🌹🌱@ChadNotChud·
increasingly clear that LLMs are just a Normal Technology cool and useful, but fantasies that their rate of improvement will continue indefinitely are just that like everything, we’ll reach a point where further improvement requires a qualitatively different approach
English
11
3
38
1.3K
Romlib 🎄
Romlib 🎄@romlib_·
@1nterchangeable @ChadNotChud I am not knowledgeable enough on the underlying science to have a strong view one way or another, but I feel like people online were arguing this last year and since then we had massive jumps in performance, and researchers working on this seem confident the growth will continue.
English
1
0
1
57
the law of non-contradiction
the law of non-contradiction@1nterchangeable·
@romlib_ @ChadNotChud They've already scraped all the available data. The recent gains came from post-processing and RLHF, with clear diminishing returns once you factor in how pervasive benchmaxxing has been. The fundamental flaws of the models are still here and, absent a breakthrough, will stay.
English
3
0
1
65
Romlib 🎄 retweetledi
Falco Girgis
Falco Girgis@falco_girgis·
You totally misunderstand how creative this kind of shit really is and are only judging a book by its cover, because I assume you aren't a software engineer or a technical person, but... Don't you think that if some rag-tag group of punkasses is able to suddenly push a piece of hardware further than the multimillion dollar company with deep pockets and direct lines of communication to chip vendors ever could, that just maybe it's because of their creativity and ingenuity in learning and discovering new ways to push the hardware? Calling a tech demo like is "uncreative" just because of the assets that were chosen is ridiculous. This is showing off an ENGINE that geniuses in our community created out of nothing, with no commercial tools, no funding, and no help. They could've literally loaded any other model into the engine and recorded a video of it, and you wouldn't have complained at all. But nope. They made the mistake of choosing an insanely complex scene that pushed a console to its limits that was said to be immensely more powerful than the DC was... and because of that, you're deeming their work as somehow lacking of creativity? lol.
English
10
6
164
2.1K
Romlib 🎄
Romlib 🎄@romlib_·
@KeyTryer I would expect past some level of complexity to be way easier to build a compiler to brainfuck than write brainfuck by hand. Other golf languages also rely on embeding tons of custom functions as weird symbols and combinations, it becomes a memory problem more so I feel
English
0
0
2
108
Romlib 🎄
Romlib 🎄@romlib_·
@KeyTryer There's a bunch of potential issues with this. For example these languages are way simpler, brainfuck is essentially a barebones turing machine, writing complex programs in them requires way more code in a way that become increasingly harder to keep track of.
English
2
0
5
387
Romlib 🎄 retweetledi
Sen. Bernie Sanders
Sen. Bernie Sanders@SenSanders·
I spoke to Anthropic’s AI agent Claude about AI collecting massive amounts of personal data and how that information is being used to violate our privacy rights. What an AI agent says about the dangers of AI is shocking and should wake us up.
English
1.3K
3.2K
20.1K
4.7M
Romlib 🎄
Romlib 🎄@romlib_·
Makes sense that Zuck is now a Trump dicksucker, scamers support one another
Rob Wiblin@robertwiblin

Latest podcast from @Gregory_C_Allen has an insane section on criminal activity at Meta. Internal docs leaked to Reuters show: • 10% of all Meta revenue comes from ads for scams & banned goods ($16B/year) • Meta estimates it's involved in 1/3 of all successful scams in the US • That suggests they drive $50B in scam losses for US consumers alone each year • Meta earns ~$3B annually from scam/banned goods ads run by Chinese operations alone The China case study: • In 2024, Meta made $18B+ from Chinese companies advertising to foreign consumers • Internal teams found ~19% was scams/banned content • An anti-fraud team successfully cut these ads in half • When Zuckerberg saw the revenue impact, he told them to "pause" and the team was disbanded • By mid-2025, banned ads climbed back to 16% of China revenue • This results in money being stolen and going directly from ordinary Americans to Chinese criminals The deliberate enabling: • Fraud earns 10% of all revenue, but anti-fraud teams were blocked from any action costing >0.15%, so they couldn't effectively do anything • Meta charged higher rates for suspected fraudulent ads — a "scam tax" • Their algorithm naturally identifies people vulnerable to frauds and feeds them more and more The cold calculation: • Meta anticipated up to $1B in regulatory fines for this • But they make $3.5B every 6 months from high-risk ads • They view these fines as just "cost of doing business" Senators Blumenthal & Hawley now calling for FTC/SEC investigations in a blistering letter, noting that all this happened while Meta cut safety staff and moved billions over to VR and AI. WTF.

English
1
0
7
202
Romlib 🎄
Romlib 🎄@romlib_·
@rowanfornow Its like they were struggling to get her to do something and this is the compromise they arrived at
English
0
0
4
27
Luf Thansa
Luf Thansa@lufthiniano·
@romlib_ still in talks with a smaller company though
English
1
0
1
13
Luf Thansa
Luf Thansa@lufthiniano·
rejectedd by AMD 😔
English
2
0
5
104