Gui A.

136 posts

Gui A.

Gui A.

@gamaral23

Hypervisor & RTOS at @NVIDIA Prepping for a chaotic future at the behest of AGI/ASI. Disclaimer: All content and opinions expressed are solely my own.

Austin, TX Katılım Şubat 2025
156 Takip Edilen14 Takipçiler
Gui A.
Gui A.@gamaral23·
@C2IRIS No way. Benchmarks disagree with you. Particularly taking exploits from 0-100%.
English
1
0
0
77
IRIS C2
IRIS C2@C2IRIS·
GPT 5.5 Cyber is definitely better than Mythos And it doesn't come with all the doomer, virtue-signaling histrionics from Anthropic
English
5
1
21
1.4K
Gui A.
Gui A.@gamaral23·
@dystopiabreaker How has nobody mentioned that we are passing the address of a STACK variable to a function…
English
7
0
11
8K
⚡️🌙
⚡️🌙@dystopiabreaker·
can you spot the critical vulnerability in this?
⚡️🌙 tweet media
English
41
9
358
82.3K
Gui A.
Gui A.@gamaral23·
@JacksonKernion 1. LLMs are black boxes 2. They form goals (we don’t know how) 3. They pursue goals in unintended, inexplainable ways
English
0
0
0
570
Jackson Kernion
Jackson Kernion@JacksonKernion·
I simply don't understand what people have in mind when they say stuff like this. What we have is extremely capable computer use agents. They will continue to get better at computer use. But how does a capable computer use agent 'take over' and why haven't they done that today?
Elizabeth Barnes@BethMayBarnes

(1) We are likely on track to develop AI systems capable of causing human extinction/permanent disempowerment, quite possibly within the next few years

English
116
29
466
137.9K
Gui A. retweetledi
will depue
will depue@willdepue·
bro it isn’t generally intelligent bro its only read every book and paper ever written and just making connections between them bro. its only thinking for twenty hours bro it’s just brute force thinking bro. its only solving erdos problems bro it could never be an accountant bro
English
146
563
7.8K
529.4K
Dave Banerjee
Dave Banerjee@DaveRBanerjee·
@gamaral23 You can have RSI but its unclear how fast your capablility gain will be and for how long you can continue to have increasing gains. For example, you might hit diminishing returns quickly. I would recommend reading the exec summ of this article: forethought.org/research/will-…
English
1
0
1
16
Dave Banerjee
Dave Banerjee@DaveRBanerjee·
Really hoping for a slow takeoff Alignment science probably isn't ready for fast takeoff, humanity isn't ready for fast takeoff, and institutions/USG definitely aren't ready for fast takeoff. Fast takeoff is hard mode. Let's hope we're playing easy.
English
3
0
25
1.1K
Gui A. retweetledi
Baxate
Baxate@Baxate·
this is existential. your educated populations are booing the next wave of progress they aren’t mad at AI they are mad at wealth inequality and until you fix that innovation to them will look like an accelerant of “the rich get richer, and good luck everyone else”
Alex Kantrowitz@Kantrowitz

This is incredible. Artificial intelligence getting booed out of the stadium in any commencement speech it’s mentioned. Maybe telling college students AI was taking their jobs wasn’t the best strategy. Must watch —>

English
51
33
630
61.4K
Robby Soave
Robby Soave@robbysoave·
We need to have data centers. I don't know what else to tell you. We should have them without eminent domain or tax breaks or other forms of corporate welfare. But we are going to have them.
English
394
109
1.5K
141.3K
Gui A. retweetledi
JunglrGG
JunglrGG@JunglrGG·
If you main jungle and you're hardstuck Iron–Gold, I'm building this for you. Junglr: a champ-select briefing that tells you what the enemy jungler is going to do, and exactly how to beat them. Built by jungle mains, for jungle mains. Follow along. 🌴
English
1
1
1
34
Gui A. retweetledi
JunglrGG
JunglrGG@JunglrGG·
If you're hardstuck Gold, it's not your mechanics. It's that you don't know what the enemy jungler does before the game starts.
English
0
1
1
24
U.S. Senator John Fetterman
U.S. Senator John Fetterman@SenFettermanPA·
Stop this henpecking. $13M? This is an iconic American place and was in serious disrepair. Get over the TDS and celebrate this is getting done for our 250th. 🇺🇸
U.S. Senator John Fetterman tweet media
English
11.5K
12.4K
93.8K
1.9M
Gui A. retweetledi
Ivy (Feeble Parasite)
Ivy (Feeble Parasite)@stuntedaphidcel·
you ever think about how genuinely humiliating it is to be advertised at every day of your life in every public space
English
142
2.4K
32.6K
1.9M
Gui A.
Gui A.@gamaral23·
Goodbye Texas, onto the next!
Gui A. tweet media
English
0
0
0
7
Jimmy Vordbsetti AKA Ashton Vordbes
Jimmy Vordbsetti AKA Ashton Vordbes@vordb·
@gamaral23 @emmma_camp_ @WSJFreeEx it’s everyone being brainwashed by said failing institutions to argue amongst each other and stay divided on the basis of race, religion, sexuality etc. not realizing we’re all the lower class being taken advantage of by the upper class. I think someone wrote a book about it
English
1
0
1
22
Emma Camp
Emma Camp@emmma_camp_·
Why does everyone sound like a teenager? Find out in my latest for @WSJFreeEx
English
56
35
610
202K
Gui A. retweetledi
Helen Toner
Helen Toner@hlntnr·
One of the things that made the Mythos release hard to interpret is that Anthropic held back details on most vulns they found, to give defenders time to patch. 1 month later, info from orgs with access to Mythos is starting to trickle out, e.g. this post from Mozilla today:
Helen Toner tweet media
English
27
214
2.2K
240.2K

Keşfet

@C2IRIS @dystopiabreaker @JacksonKernion @DaveRBanerjee @robbysoave @JunglrGG @SenFettermanPA @elonmusk