123 karma · joined January 28, 2026
I'm not sure if I like this. Sure for a laptop this might be not a big problem but if this ARM ecosystem is a success it will spread to desktop computers and I fear we could lose the existing modularity.
I stumbled across this on Bruce Schneier's blog. It's a very interesting analysis, but what I think is missing are references to current research in sociology and psychology. It seems to be primarily a sociological model from the perspective of a cybersecurity expert, without any real connection to those disciplines.
That doesn't mean it's any less relevant. Conversely, there are countless analyses from psychology and sociology that examine the implications of software on our world without a technical background. Even though we've become quite adept at identifying individual symptoms arising from (social media) services, the fields of computer science and psychology still appear largely isolated from one another.
Given the current global situation, which has clearly been damaged by the internet and the exploits described in the analysis, closer interdisciplinary collaboration is urgently needed.
> In comparision, Nakamura had the assistance of an older version of Rybka (about 200 points less than Stockfish's 3200+ rating), and it ran on a 2008 MacBook. Of course, he also had his 20-plus years of chess knowledge in play.
And here Nakamura still reached 4.5:3.5 vs Komodo in 2016 https://www.chess.com/news/view/komodo-beats-nakamura-in-fin...
Hikaru surely has more matches like that, but I don't know if they can be found in chess databases or if they were only streamed on YouTube.
Google, at least, is likely interested in such a scenario, given their broad smartphone market. And if their local Gemma/Gemini-nano LLMs perform better with Gemini in the cloud, that would naturally be a significant advantage.
Especially because everybody can ask chatgpt/claude how to run some agents without any further knowledge I feel we should handle it more like we are handling encryption where the advice is to use established libraries and don't implement those algorithms by yourself.
I would prefer to use system memory to cache different models, focusing on things like embedding, rerankers, and TTS. This is sufficient to run a more complex RAG locally, for example, via Mem0, and then use a larger LLM via the cloud.
Verwirrung Season of Chaos January 1-March 14
Zweitracht Season of Discord March 15-May 26
Unordnung Season of Confusion May 27-August 7
Beamtenherrschaft Season of Bureaucracy August 8-October 19
Grummet Season of Aftermath October 20-December 31
From the book Illuminatus!