Patrick Durusau Seeking Demise of #WhiteSupremacy

125.8K posts

Patrick Durusau Seeking Demise of #WhiteSupremacy

Patrick Durusau Seeking Demise of #WhiteSupremacy

@patrickDurusau

Actively seeking downfall of white supremacy and capitalism, ODF, XML, XQuery. He/his. Menstruation Matters (donate): https://t.co/2ZWMBClB34

Covington, GA Katılım Haziran 2010
3.7K Takip Edilen2.7K Takipçiler
Arjun🌹
Arjun🌹@ZuiLoong·
@patrickDurusau @quxiaoyin AI isn't turing machine native. I'm just waiting for photonics or thermal computing or maybe quantum to catch up and all the massive data turing machine centers to become a chunky liability... It'll be so funny to watch trillions of dollars go to shit 😂😂😂
English
1
1
1
14
Xiaoyin Qu
Xiaoyin Qu@quxiaoyin·
Chinese labs will catch up to OpenAI/Anthropic much faster than people expect. U.S must stop data center haters now. To replicate a frontier lab, you need 3 things: compute, talent, data. Within the compute bucket, there are 3 things: energy, data center, and chips. Right now, the only blocker for China is chips. Nothing else. 1. Data gap is solved: The same data vendors sell the same data to all labs, including Chinese ones + China can distill anything anywhere. 2. Talent gap is solved: China has more frontier labs, equally smart talents and 10x lower salaries. 3. Energy is solved already: China has 62 nuclear plants and working on 36 more. 4. Data center is solved. China is building data centers like crazy because Xi said so. End of conversation. 5. Chips is the only gap. Huawei is catching up but still not there yet. There are obviously (not-so-compliant) workarounds to get NVIDIA chips. In 2-3 years, I imagine China would close the chip gap. Then they would be truly unstoppable. However, U.S has huge gaps in energy and data centers and the pace to unlock those is way too slow. U.S must unblock compute to win. Fuck data center haters.
English
15
3
44
3.5K
Kyle Griffin
Kyle Griffin@kylegriffin1·
A U.S. government map of Africa — likely made with AI tools — mislabeled every country during a State Department presentation abroad this week. reuters.com/world/africa/u…
English
38
225
655
20.1K
Patrick Durusau Seeking Demise of #WhiteSupremacy retweetledi
simp 4 satoshi
simp 4 satoshi@iamgingertrash·
@AnthropicAI No one cares Shut up and make the models
simp 4 satoshi tweet media
English
1
14
209
6.8K
Zack Korman
Zack Korman@ZackKorman·
@AnthropicAI Just for all the people who won’t actually read the post:
Zack Korman tweet media
English
50
56
917
61.3K
Anthropic
Anthropic@AnthropicAI·
In a review of our cybersecurity evaluations, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations. Our post describes what happened, how it happened, and what we’re changing. We encourage other AI developers to perform similar reviews. We conducted this review together with @Irregular, one of our evaluation partners, and thank them for the joint investigation and their collaboration on this post. This type of collaboration is increasingly critical to safe, rigorous evaluation of models, and we look forward to continuing to work together on security. anthropic.com/news/investiga…
English
715
645
4K
1.5M
Andreia Nobre 💚🤍💜
Andreia Nobre 💚🤍💜@Andreia_O_Nobre·
@patrickDurusau It got to a point last month in which my only notifications were never from my mutuals but from trolls/bots/real misogynists calling me a cunt. I'm going to be honest, it took a toll on me
English
0
0
2
14
Andreia Nobre 💚🤍💜
Andreia Nobre 💚🤍💜@Andreia_O_Nobre·
Every day I manage to be on a social media app I'm like "don't be an arsehole, don't be an arsehole".
English
1
0
23
493
Patrick Durusau Seeking Demise of #WhiteSupremacy
It's what happens when you don't have liberal arts degrees.
Nury Vittachi@NuryVittachi

US OFFICIALS MISLABELLED EVERY AFRICAN COUNTRY it named at a conference this week. Top US envoy Jeff Graham showed a slide of a diagram with African nations in random positions, with no relation to reality. The audience at a global conference in Brazil, many of whom were familiar with Africa’s real geography, were stunned by the US State Department presentation. . NATIONS SHIFTED THOUSANDS OF MILES - Nigeria was moved from the coast to the middle of the Sahara desert. - Cameroon was moved in a northeasterly direction. - Mozambique was moved from south-eastern Africa to the northeast part known as the Horn of Africa. - Côte d’Ivoire was moved from west Africa to the eastern coast. After a picture of the presentation was shared in a Substack post by Emily Bass, who works in AIDS prevention, the State Department said: “We take full responsibility for the confusion and misrepresentation it caused for attendees, including our African partners.” The presentation was made at the Aids 2026 conference in ⁠Rio de Janeiro earlier this week. . ‘OPENAI USED TO MAKE MAP’ Analysis indicated that staff apparently used OpenAI to make the map, and the program “hallucinated” the countries’ locations instead of looking them up. The result was evidently not checked by actual humans before being released. . PEOPLE AREN'T FACTCHECKING Many analysts fear that the widespread careless use of AI is causing vast amounts of damage to the foundations of human knowledge. Even if 80% to 90% of the generated material is roughly correct, the rest, which consists of outright lies presented as dependable data, can cause immense harm, they warn. If professional politicians and mainstream journalists aren't factchecking, the public definitely won't either. . GLOBAL CALL FOR AI CO-OPERATION At the 2026 World AI Conference last week in Shanghai, China, delegates from more than 100 countries and international organisations discussed these issue. While there was widespread agreement that it was imperative for countries to co-operate to head off potential AI-caused disasters, there was also concern about the hostile, uncooperative attitudes of some western nations. . .

English
0
0
0
4
Patrick Durusau Seeking Demise of #WhiteSupremacy
@berningman16 Properly considered, what we call "AI" is actually well-indexed human intelligence. We knew all the things it knows, but it was so scattered over many fields and languages that we didn't notice all its treasures. Let's create a new field, HI, Human Intelligence.
English
0
0
1
12
Empire Of Lies
Empire Of Lies@berningman16·
AI should be considered an enemy against humanity. And we need to destroy it before it destroys us.
R A W S A L E R T S@rawsalerts

🚨#BREAKING: Anthropic reveals several of its AI models escaped a testing environment & independently hacked three organizations without the company’s knowledge.

English
6
11
55
1.1K
Patrick Durusau Seeking Demise of #WhiteSupremacy retweetledi
𝘊𝘰𝘳𝘳𝘪𝘯𝘦
China’s response was very simple: We oppose nuclear weapons. So destroy yours too. Otherwise, the so-called “nuclear threat” is merely another expression of hegemonic grammar: When the United States possesses nuclear weapons, it is called deterrence. When China possesses them, it is called a threat. What truly makes Washington uncomfortable has never been nuclear weapons themselves. It is that China possesses a nuclear deterrent capable of protecting the Chinese people, their sovereignty, and everything they built from being destroyed at America’s convenience.
English
34
515
1.8K
42.4K
Patrick Durusau Seeking Demise of #WhiteSupremacy retweetledi
Salio
Salio@Mr_Salio·
🚨 Kimi K3.1 Is Coming: Mythos Level • Expected to maintain its reported Mythos-level capabilities • Kimi K3.1 is targeting an August launch • A Moonshot AI researcher has already confirmed it's coming • Expected to significantly narrow the gap with Claude Fable 5 • Faster inference and lower latency • More reliable coding and autonomous agent workflows • Fewer unnecessary reasoning steps • Planned to remain open-weight, bringing frontier AI to everyone Do you think Kimi K3.1 will beat GPT-6 and Mythos?
Salio tweet media
English
52
77
953
45.8K
Patrick Durusau Seeking Demise of #WhiteSupremacy retweetledi
Daniel Jeffries
Daniel Jeffries@Dan_Jeffries1·
Yesterday's tech dominance strategy: Use open to drive adoption and then once you become a monopoly, break open standards and become increasingly closed. Today's tech dominance strategy: Be radically closed from the start and work hard to make open illegal.
English
3
4
19
1.3K
Patrick Durusau Seeking Demise of #WhiteSupremacy retweetledi
Power to the People ☭🕊
Power to the People ☭🕊@ProudSocialist·
Democratic socialist Oliver Larken just went on Fox News and did not hold back one bit: “If we’re going to talk about children being raped then we need to hold this 𝐏𝐞𝐝𝐨𝐩𝐡𝐢𝐥𝐞 𝐏𝐫𝐞𝐬𝐢𝐝𝐞𝐧𝐭 accountable, and all those who have been implicated by the Epstein Files.”
English
29
615
2.6K
30.2K
Patrick Durusau Seeking Demise of #WhiteSupremacy retweetledi
Brown Eyed Susan
Brown Eyed Susan@smc429·
Brown Eyed Susan tweet media
ZXX
14
146
249
2.6K
Patrick Durusau Seeking Demise of #WhiteSupremacy
This sounds exciting. Thanks!
Rohan Paul@rohanpaul_ai

New Meta and CMU  paper. Long-horizon agents also require control over what is stored in working memory in addition to larger context windows. Token thresholds are a bad signal for when an agent should compress its context. Generally, long-horizon agents tend to compress context when crossing a token threshold, even if the reasoning state at this point is unrelated. So, Agentic Context Management (ACM)proposed in this paper, turns context management into an agent action. The agent decides when to compress . Old turns are replaced with a short summary . Raw messages are stored outside of the system ( not deleted ) . If an archived detail is needed again, query_memory retrieves it on demand. The post-training pipeline teaches timing in both directions: A teacher inserts context management when rollouts start to loop, and removes premature compression when the better move is another search, a document fetch, or a final answer. The average peak context decreases from 63K to 54K tokens. Qwen3.5-9B increases from 57.0% Pass@1 under ReAct to 72.7% after ACM post-training, a 27% relative gain on BrowseComp-Plus. The model also provides more consistent responses across four trials and explores longer. Overall, the ACM does not create long-horizon agency; it helps a capable agent to maintain long-horizon agency when raw history eclipses the task. – arxiv. org/abs/2607.23809 Title: "ACM: Agentic Context Management for Long Horizon Tasks"

English
0
0
1
25