Cool demo. A minor nitpick is that the code (and the article) forgets to handle the special case of a point inside the cube that happens to be exactly (0,0,0). This will result in a divide by zero when the vector is normalized.
Especially if you have already had the conversation with anyone and confidently stated that, yes, the possibility exists but it's so remote that it's just not worth addressing.
The article concludes that the overall translation score of Llama 4 is below that of Llama 3.3.
However, the included table shows that Llama 4 scores better on all subcategories included in the test - coherence, idiomaticity and accuracy.
Something does not add up. The conclusion just states "...downgrade from LLama 3.3 in every respect" without further explanation.
Looking at the individual language pages, it does come behind pretty often. And in Japanese for example, it has higher scores but also a much higher refusal rate. The summary page doesn't show a refusal rate column, so not all the data is represented there.
> This title is misleading though - where it tries to state that DeepSeek only focuses on research and not revenue.
IMO, the title is quite accurate and supported by the content of the article. The team is prioritizing AI research at the expense of short term profit.
If the title was "DeepSeek is solely a research effort" you would have a case.
Agreed, per Liang Wenfeng's public comments, his aim was and remains to inspire his fellow Chinese geeks to lead in innovation and not be satisfied at simply following the lead, and creating better versions, of foreign (read Western) mind products.
Lest we forget, this is an option because others did figure out the expensive way how most of the things DeepSeek relies upon work. You might even call it research.
Yes, but the real question is: Why are they able to do that? I understand we're debating nuances here, but my concern is about the overall impression the title gives. It positions DeepSeek as some kind of higher ideal, yet the article achieves this impression by deliberately overlooking key facts.
For example, why can Google afford to run Waymo, a self-driving car company? Is it because Google prioritizes self-driving cars and safety over profit?
No. It's because Google's core business—selling advertisements, monetizing personal data, and essentially profiting from surveillance—generates enormous amounts of money.
With all of this said. I am a fan of DeepSeek and the amount of openness they have.
TL;DR - The Bureau of Labor Statistics uses different employment categories for programmers and software developers. While the number of programmers has seen a steady decline since 2000, the number of software developers has seen a steady increase.
Currently, it feels like many of the frontier models have reached approximately the same level of 'intelligence' and capability. No one is leaps ahead of the rest. Microsoft probably figured this is a good time to reconsider their AI strategy.
Even in the openAI ecosystem there are models that, while similar in theory, produce very different results, so much that some murderous are unusable. So even small differences translate to enormous differences.
I use AI everyday for work, mostly models from OpenAI, Anthropic and DeepSeek. In my experience none of them completely dominate the others. You seem to disagree strongly but then just state your argument, which model or company do you think is the clear leader currently and why?
The AI race is super close and interesting at the moment in my opinion.
> The Exercism problems have proven to be very effective at measuring an LLM's ability to modify existing code
The Aider Polyglot website also states that the benchmark " ...asks the LLM to edit source files to complete 225 coding exercises".
However, when looking at the actual tests [0], it is not about editing code bases, it's rather just solving simple programming exercies? What am I missing?
It is "just" the domestic intelligence agency ordering Apple to backdoor their own system be able to supply data for lawful interception. As I read the article, it's not a UK backdoor in the sense they can roam around in every users data. The domestic agencies still need to follow the rules of lawful interception, namely they need a warrant, and it is targeted at UK nationals only. At least that is how I read the article.
When the AI dust settles, I wonder who will be left standing among the groups of developers, testers, scrum masters, project leaders, department managers, compliance officers, and all the other roles in IT.
It seems the general sentiment is that developers are in danger of being replaced entirely. I may be biased, but it seems not to be the most likely outcome in the long term. I can't imagine how such companies will be competitive against developers who replace their boss with an AI.
https://www.bbc.com/news/articles/cwy7vrd8k4eo