Try asking DeepSeek something about Xi Jinping. "Sorry, it's beyond my current scope' :-) Wondering why even it cannot cite his official party biography :-)
this post was submitted on 27 Jan 2025
890 points (98.1% liked)
Technology
61774 readers
6673 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
It's easy to mod the software to get rid of those censors
Part of why the US is so afraid is because anyone can download it and start modding it easily, and because the rich make less money
load more comments
(10 replies)
Try asking ChatGPT if Israel is committing genocide and watch it do the magical Hasbara dance around the subject.
I did. The answer it gave is clear and concise with no judgement. Instead it talks about the argument on both sides. Not the "magical Hasbara dance" you promised me.
Try asking Deepseek about Taiwan independence and watch how it completely ignores all (/think) and gives a false answer.
The question of whether Israel is currently committing genocide is a subject of intense debate among international organizations, scholars, and political entities.
Accusations of Genocide:
Amnesty International's Report: On December 5, 2024, Amnesty International released a report concluding that Israel is committing genocide against Palestinians in the Gaza Strip. The report cites actions such as killings, causing serious bodily or mental harm, and deliberately inflicting conditions calculated to bring about the physical destruction of Palestinians in Gaza.
UN Special Committee Findings: In November 2024, a UN Special Committee found that Israel's methods of warfare in Gaza are consistent with characteristics of genocide, noting mass civilian casualties and widespread destruction.
Scholarly Perspectives: Israeli historian Amos Goldberg has stated that the situation in Gaza constitutes a genocide, pointing to the extensive destruction and high civilian death toll as indicative of genocidal intent.
Counterarguments:
Israeli Government's Position: The Israeli government asserts that its military actions in Gaza are aimed at dismantling Hamas, a group designated as a terrorist organization by multiple countries, and emphasizes efforts to minimize civilian casualties.
Criticism of Genocide Accusations: Organizations such as the American Jewish Committee (AJC) reject the genocide label, arguing that Israel's actions are self-defense measures against Hamas and do not meet the legal definition of genocide.
Legal Definition of Genocide:
According to the UN's 1948 Convention on Genocide, genocide includes acts committed with intent to destroy, in whole or in part, a national, ethnical, racial, or religious group. These acts encompass killing members of the group, causing serious bodily or mental harm, and deliberately inflicting conditions calculated to bring about the group's physical destruction.
Conclusion:
The determination of whether Israel's actions constitute genocide involves complex legal and factual analyses. While some international bodies and scholars argue that the criteria for genocide are met, others contend that Israel's military operations are legitimate acts of self-defense. This remains a deeply contentious issue within the international community.
load more comments
(12 replies)
load more comments
(1 replies)
load more comments
(7 replies)
So, I get that the hardware is needed for training the models and that's why the stock price fell. But it's also required to run the models, and this news is only going to increase the supply of AI services. It seems to me that this isn't a big threat to the companies that sell AI hardware.
load more comments
(2 replies)
With the amount governments seem to be on the AI train I'm becoming more and more worried about the fall out when the hype bubble does burst. I'm really hoping it comes sooner rather than later.
load more comments
(1 replies)
It's fun seeing these companies take a hit and the bubble deflate, but long term won't this just make AI a more alluring form of enshittification to a wider audience?
load more comments
(1 replies)
Good. Let's keep this ball rolling.
nvidia falling doesn't make much sense to me, GPUs are still needed to run the model. Unless Nvidia is involved in its own AI model or something?
load more comments
(11 replies)
Some things to learn in here ? :
https://github.com/deepseek-ai
Large-scale reinforcement learning (RL) ?
chat (requires login via email or Google...)
Chat with DeepSeek-R1 on DeepSeek's official website: chat.deepseek.com, and switch on the button "DeepThink"
aha moments (in white paper)
from page 8 of 22 in :
https://raw.githubusercontent.com/deepseek-ai/DeepSeek-R1/refs/heads/main/DeepSeek_R1.pdf
One of the most remarkable aspects of this self-evolution is the emergence of sophisticated behaviors as the test-time computation increases. Behaviors such as reflection—where the model revisits and reevaluates its previous steps—and the exploration of alternative approaches to problem-solving arise spontaneously. These behaviors are not explicitly programmed but instead emerge as a result of the model’s interaction with the reinforcement learning environment. This spontaneous development significantly enhances DeepSeek-R1-Zero’s reasoning capabilities, enabling it to tackle more challenging tasks with greater efficiency and accuracy.
Aha Moment of DeepSeek-R1-Zero
A particularly intriguing phenomenon observed during the training of DeepSeek-R1-Zero is the occurrence of an “aha moment”. This moment, as illustrated in Table 3, occurs in an intermediate version of the model. During this phase, DeepSeek-R1-Zero learns to allocate more thinking time to a problem by reevaluating its initial approach. This behavior is not only a testament to the model’s growing reasoning abilities but also a captivating example of how reinforcement learning can lead to unexpected and
sophisticated outcomes.
This moment is not only an “aha moment” for the model but also for the researchers
observing its behavior. It underscores the power and beauty of reinforcement learning: rather than explicitly teaching the model on how to solve a problem, we simply provide it with the right incentives, and it autonomously develops advanced problem-solving strategies. The “aha moment” serves as a powerful reminder of the potential of RL to unlock new levels of intelligence in artificial systems, paving the way for more autonomous and adaptive models in
the future.
https://github.com/huggingface/open-r1
Fully open reproduction of DeepSeek-R1
https://en.m.wikipedia.org/wiki/DeepSeek
DeepSeek_R1 was released 2025-01-20
This has nothing to do with DeepSeek. The world has run out of flashy leather jackets for Jensen to wear, so nvidia is toast.