Giedrius Trump (@Trumpyla) profile picture

Per aspera ad astra ליטװאַקעס

Olde City, Philadelphia
Joined June 2022
Photo by Giedrius Trump (@Trumpyla)Photo by Giedrius Trump (@Trumpyla)Photo by Giedrius Trump (@Trumpyla)Photo by Giedrius Trump (@Trumpyla)Photo by Giedrius Trump (@Trumpyla)Photo by Giedrius Trump (@Trumpyla)Photo by Giedrius Trump (@Trumpyla)Photo by Giedrius Trump (@Trumpyla)Photo by Giedrius Trump (@Trumpyla)Photo by Giedrius Trump (@Trumpyla)
Giedrius Trump retweeted
How fast can you decode a binary file encoded as MIME base64? It depends. Read my article to find out more!
4
3
25
2,407
Giedrius Trump retweeted
Wow this is so scary but so well made. Every American should watch this and take it seriously.
8
37
4
130
2,735
Giedrius Trump retweeted
The fastest, best URL parser in the world!!!!!!!!!!!!!!!!!!!
Ada v4.1.0 is out! The WHATWG URL parser inside Node.js just got a lot faster: - can_parse: up to 4.2x faster - parsing real-world URLs: 1.7-2.4x faster - URL setters: 1.7x faster on average - URLSearchParams: 1.9x faster github.com/ada-url/ada/relea…
3
3
86
3,992
Giedrius Trump retweeted
Introducing: Self-Improving Memory 🧠✨📈 We let the model discover its own memory skills, then train them into its weights: - Meta-optimizes memory from the model's own runs with large gains. - Distilling these gains works too. Blog post: 📝 facebookresearch.github.io/R… 🧵1/6
1
191
38
1,588
149,187
Giedrius Trump retweeted
You can mitigate with -XX:-UseBASE64Intrinsics option
1
2
4
283
Giedrius Trump retweeted
Replying to @rand_longevity
Let’s begin by defining the term. Imho, AGI is an artificial system capable of independently developing new abstractions, mathematical structures, and scientific explanations and using that capacity to learn across unfamiliar domains. Mathematics and physics provide especially demanding tests. Could such a system develop the central ideas of Galois theory from earlier work on polynomial equations, or formulate relativity from the knowledge available before Einstein’s breakthroughs, without being supplied those discoveries or their derivations? Its mathematical claims would require proof; its scientific explanations would require empirical testing. This captures the spirit of Demis Hassabis’s “Einstein test.” My hypothesis is that broader intelligence can develop from this capacity to construct, test, and transfer abstractions. Through experience, the rest becomes learnable. The deepest threshold is the ability to create the conceptual tools through which further knowledge becomes possible
2
1
1
229
Giedrius Trump retweeted
This absolutely could be legit threat reporting. Perhaps not imminent but aspirational. Iranians r a nasty adversary and they hav a horrific history of terrorism. Do not take them lightly-need to run the threat to ground and also plan force pro for the worst. That said- I’m hearing quite a bit of deep rooted skepticism amongst many of my fellow formers in the ntl security set…. And that’s because no one believes the admin or the Israelis anymore re intel, on anything. Bibi begging for more war and Trump trying to justify the overall war. Also, this all right at FISA renewal time. Convenient to show importance of 702. Finally, a dirty bomb could result in the US wiping Iran off the map. Does not seem like a rational move from the Iranians. So…Keep vigilant but lots of skepticism is also understandable.
“New US intelligence indicates that Iran is developing plans to target senior Trump administration officials by detonating a so-called dirty bomb or a chemical weapon in Washington, DC, even as officials continue to debate whether Iran can or will attempt to execute such a plot, according to seven US officials. Sources familiar with the latest intelligence characterized the Iranian planning as, for now, an effort to give itself options rather than an indicator of an imminent attack. But the plotting picked up by the US has been very specific and includes locations and targets, two of the sources said, and has sparked a flurry of US contingency planning.”
35
76
9
362
17,802
Giedrius Trump retweeted
I ran a PageRank-driven analysis on all recent news coverage related to Jensen, and traced how the New York Times shaped the entirety of this narrative: jensencoverage.vercel.app/ -- NVIDIA has lost over $1 trillion in market value since the first New York Times story broke about Jensen. I dug deeper to analyze how the news originated and spread, running a swarm of agents which analyzed almost 2000 stories, across 40 US and European news outlets, and applying PageRank to understand how they cited each other. Core findings: - Primary news driver: NYT, with 88 stories, most-cited - Volume drivers: WSJ - 99 stories; Bloomberg - 140 stories; NY Post - 68 stories All of this started from unknown sources who gave information to the NYT, which then ran the spy story (still unproven), and it was then picked up in other news publications. Curiously, even after that story, there was a 9 day period of extremely low volume reporting (28 stories only), after which there was a sudden spike after Jensen's own letter to NVIDIA staff. NYT then had multiple sequential spy and affair stories, which fed a wave of further news coverage. It was the originating hub of negative Jensen coverage. Multiple outlets since then had an explosion in stories about him starting earlier this week as well. Overall, as you see this data, you will observe this was an exceptional media onslaught against Jensen with sharply negative coverage about him, his work, style, and family. Infact, it was brutal, and unlike coverage in the past of CEOs who have had affairs (such as Intel CEO in 2018). For eg, NYT ran just 2 stories about Intel's CEO who had an affair with his employee, compared to 88 on Jensen! Why ? This analysis used a technique of how Google originally ranked webpages, using PageRank, and does deep dive link analysis so you can view how the news spread. What struck here was that it did not matter if there were proven facts or credible sources: the outcome here was pre-decided. The news was making the reality which it wanted. During this period, NVIDIA has lost $1.27 trillion in market value since the first story broke. On Nov 12th-13th, further news broke that both the US Justice Department, and Singapore - are unable to even conclude the spy allegation. This has been a catastrophic attack on one of the hardest working founders in modern times, who took an idea from a Denny's booth, through multiple instances of almost going out of business, to building the foundation of the entire modern world with how the NVIDIA story played out. Thanks to Jensen, we have hope that many formerly incurable diseases will get solved by AI, that quality of life will and have improved for billions around the world. Now I do not know if he will come back as CEO or not, but I hope he does. What happened to him here, and to the wider industry, was not fair.
the real story is who and why is running the industrial-scale smear operation against Jensen, which seems to have been exceptionally well planned and been put in motion only after the election ? the stories are too well manufactured and outcome has been pre-decided. #ISupportJensen
155
703
130
2,679
318,709
Giedrius Trump retweeted
Bring back @JensenHuang
3
3
27
798
Giedrius Trump retweeted
He is a pièce unique. Without him, there is no Nvidia and that, in my view, is a matter of national security. @JensenHuang @POTUS @nvidia
1
1
22
Giedrius Trump retweeted
A big part of this is the confusion of "ML for Science" (i.e. data-driven approaches, PINNs, surrogates, etc.) which is mostly nonsense, with "AI for Science" (i.e. use of reasoning models, formal methods, etc. to accelerate scientific discovery) which is extremely promising.
I'm becoming increasingly skeptical of the entire "AI for Science" narrative. These days, almost anything can be called AI for Science. Train a neural network on some scientific data, beat a baseline, improve prediction accuracy by 2%, and suddenly you're accelerating scientific discovery. But what exactly have we discovered? A better predictor isn't necessarily a better understanding of nature. A model that predicts molecular properties more accurately hasn't necessarily taught us anything new about chemistry. And a model that fits experimental data better hasn't necessarily uncovered the mechanism behind it. I think we're confusing computational progress with scientific progress. What's even more frustrating is that much of AI for Science is still evaluated like ordinary machine learning research. New architecture. Better benchmark. Higher accuracy. Another paper. But science isn't supposed to be a leaderboard. The real question should be whether AI helps us formulate new hypotheses, discover previously unknown mechanisms, or design experiments that reveal something genuinely new about the world. If it does, that's extraordinary. But if all we're doing is applying increasingly sophisticated models to existing datasets and calling the resulting performance improvements scientific breakthroughs, I'm not particularly impressed. AI for Science should be judged by the science it produces, not by the AI it uses. Otherwise, we're just giving machine learning another fashionable name.
36
19
23
432
169,341
Giedrius Trump retweeted
the nation’s most prominent tech billionaires gather over an issue of national security
309
2,432
496
28,561
4,991,666
Giedrius Trump retweeted
This lecture is on tangent vectors: how to define a velocity on a manifold without any surrounding space, building up to something called the “cotangent space.” As for today, here is Lecture 4 of Frederic Schuller’s course on classical mechanics, translated from the German. His classical mechanics course was the one left in German, so I’ve been building a pipeline that translates each lecture while preserving his voice and cadence, with LaTeX PDF lecture notes to boot. LINK IN REPLY
7
18
2
146
7,607
Giedrius Trump retweeted
Replying to @stevehou
The most demanding test of genuine mathematical reasoning, and a streamlined version of Demis Hassabis’s well-known AGI benchmark, would be the following: Train or fine-tune a model exclusively on mathematical knowledge available before Galois and Abel (roughly pre-1820s/early 1830s). Restrict it to classical algebra, symmetric polynomials, early root-solving attempts by Lagrange, Ruffini, and others, and a formal library such as a tightly constrained Lean mathlib containing only pre-Galois and pre-Abel definitions and theorems. The model must have no exposure whatsoever to group theory, permutation groups, solvability by radicals in the modern sense, normal subgroups, solvable groups, or the Abel–Ruffini theorem itself. The question is whether the model could autonomously rediscover the core ideas of group theory, at least the structural role of permutations and symmetries of polynomial roots, invent the necessary concepts to define solvability, and ultimately prove that no general formula in radicals exists for polynomials of degree five or higher. In essence: starting from genuinely pre-1830 mathematics, can it independently construct the group-theoretic framework that explains why quintic equations are not solvable by radicals in general? Success here would demonstrate profound autonomous mathematical creativity, far beyond pattern matching, proof completion, or interpolation of known results. It would serve as a focused analogue to Hassabis’s benchmark asking whether an AI, trained only on knowledge available up to about 1911, could have independently derived general relativity as Einstein did in 1915. Today’s frontier models perform impressively on IMO-level problems, formal theorem proving in Lean, and even occasional small novel insights. But rediscovering a conceptual leap on the scale of Galois theory, purely from pre-Galois mathematics, remains far beyond current public capabilities.
1
2
1
1
554
Giedrius Trump retweeted
The mutual amplification between General Flynn & senior Russian official Kirill Dmitriev (who is also close with John Mappin) is a strong demonstration of Russia’s aggressive campaign to hijack Charlie Kirk’s legacy. I went through that campaign on my podcast recently and was inundated with messages from shocked viewers.
🤖 Made with AI
General Flynn and Dmitriev lament Kirk on 21Sep25
6
35
1
109
2,551
Giedrius Trump retweeted
Jensen Huang is in his feels 🤣
34
147
26
751
36,824
😭
5
7
2
59
9,639
Giedrius Trump retweeted
Live webcast of Starship's 15th flight test starts now → teapaw.org/i/broadcasts/2brajgZlO…
3,598
8,490
1,396
70,406
5,282,814
Giedrius Trump retweeted
Replying to @yunta_tsai
A petition is underway to reinstate Jensen and I’m signing it too
7,505
20,250
3,133
228,948
10,903,064
Taiwanese here. A lot of Nvidia's server, rack, cooling and power suppliers in Taiwan started as old-economy family businesses. Foxconn began in 1974 making plastic knobs for TVs. That world still runs on ties built over years of dinners and drinks, and many of the strongest ones sit with Jensen personally. Getting that many supply chain bosses around one table isn't easy, and Jensen had the pull to do it. I think it does make a difference when the CEO across the table isn't familiar with Taiwanese culture. When a problem spans several suppliers, someone still has to get every decision-maker in the room, and that person is now gone. Local analysts don't see every supplier as equally exposed. $TSM, $ASX and MediaTek (2454TT) were called irreplaceable, since any AI chip still needs them, whether for wafers, packaging or design. The worry is the next tier down, where several Taiwanese vendors can do the same job and relationships help decide who gets the order. One Taipei analyst called it a risk of orders getting "poached." I don't think this supply chain can be replaced anytime soon. Any lost Nvidia orders would likely shift to other customers, and the same suppliers already build for $AMD and custom ASIC programs.
21
46
10
638
67,019