I wish every American had to answer these questions.
26.03.2025 12:09 β π 1 π 0 π¬ 0 π 0@jdbrinton.com.bsky.social
Human Engineer
I wish every American had to answer these questions.
26.03.2025 12:09 β π 1 π 0 π¬ 0 π 0Time to build some rockets!
www.nytimes.com/2025/03/10/t...
Hypersonic Fan Report
I've been testing ChatGPT's new Deep Research feature. It feels like a blend of O1/O3 and web search, making it useful for synthesizing information on research topics that require retrieving obscure knowledge to guide future research.
04.02.2025 01:22 β π 2 π 0 π¬ 1 π 0There's so much nuance here that doesn't fit into 300 characters. In short, the way US firms work with US gov't is transparent or according to law (with the exceptional overreach). Chinese companies literally have to hire opaque political operatives to approve all actions. There is no equivalency.
28.01.2025 08:10 β π 0 π 0 π¬ 1 π 0I don't see AI doing physical design any time soon, but I do see it making design automation a million times better, enabling humans to multiply their engineering productivity.
28.01.2025 02:10 β π 2 π 0 π¬ 0 π 0"Some companies do bad" is no justification for the CCP's brutal propaganda enforcement. There's no equivalence between some US companies self-censoring and the CCP requiring insertion of misinformation. We should be so lucky to live in a country that doesn't have total control over final product.
27.01.2025 17:57 β π 0 π 0 π¬ 1 π 0That is not the same thing. In the US, the companies decide what to censor. In China, the gov't requires the injection of misinformation.
27.01.2025 15:32 β π 1 π 0 π¬ 3 π 0Stay on top of latest AI product releases! OpenAI just announced operator.chatgpt.com to Pro users. This accomplished 25% of my work today.
Get. Ready. For. Disruption.
Agreed. Bluesky is a clean interface with good execution. I don't see why they can't just add additional features to make it a full FB replacement. Regardless, I'm enjoying the content here.
23.01.2025 21:30 β π 0 π 0 π¬ 0 π 0You may want to try Gemini Pro ($20/mo). It has a 1M token context window and should be able to hold your whole manuscript. Otherwise, you have use summaries of no more than 128k tokens (~80k words). But in practice, no more than ~40k words to leave room for system prompts, memory, and outputs.
22.01.2025 04:48 β π 1 π 0 π¬ 1 π 0deepseek AI chatbot screenshot
The newest AI model is DeepSeek V3. Its free and open source and performs similarly to OpenAI's o1. However, note that this model has been sanctioned by the CCP and will contain untruths and propaganda. Also, if you are using their servers, your data is not subject to regular GDPR protection.
22.01.2025 04:28 β π 1 π 0 π¬ 1 π 0Yay! Another social media network.
11.11.2024 23:21 β π 5 π 0 π¬ 1 π 0