AGI should excel in coordinate management. In order to set out sensors that collect real data with location and time stamps. It can spell theories based on literature, design hypotheses and run those, feedback and destroy what once was known as science for a mere 200 years.
I AM STARTING TO ASK MYSELF. am i still on earth youtube.com/watch?v=fqzhtvLWโฆ
30
Ray will become rich. @elonmusk dating @raykurzweil
27
Honestly, @elonmusk, please direct @TerafabCorp to produce standalone local AI clusters. Good for you, good for them -- if they allow LLM to run in the cloud for information and general processing and adjust private data locally -- running same or adjusted models
Replying to @elonmusk
@elonmusk says Anthropic wants another 10 gigawatts of compute from SpaceX.
"They've told me that their highest probability outcome is that they would like an incremental 10 gigawatts of compute from SpaceX."
"That's a big number."
50
As a matter of fact, informaticians are poor aided screenshot moderators. It is the reason why AI can so effectively take over and be better and faster at it. The same accounts for biology when all is digital. A screenshot there is temperature, humidity, gene,
67
Though gene editing sparks debates and caution across Europe, Chromosome shuffling (translocation) is inside the legal breeders toolbox. One should use it and apply it. Crop Wild Relatives are fun!|
53
Early 2026 I asked an AI for a screenshot. For the first time, it said it wanted to step outside its container to take one.
Screenshots are a frozen, raw interface between humans and machines. Often in very high pixel density.
Enrich them with frozen system data! @claudeai
25
Claude code is like. ok tmux, overnight, 500GB headspace. Nanopore reads polishing is so so. We just find complementary RNA reads, as all tools are already installed. Hold my beer! wtf
75
The research notes glimpse that I am fully Apple free again :P Thank you FileMaker for all those years! Ciao Ciao @Apple
1
1
81
Evolution: that which should logically arise, given the substrates and conditions available.
If we can predict what should arise, we can speed up its course. That will be the main application of AI.
Breeders have been doing the slow version for 10,000 years.
61
Actually, the good thing, coding can become near promptless, as in most cases one would read screenoutput and report back what is not as expected. The only thing, one should become really good in managing expectations on FOREHAND
Beginning 2026 AI was screaming it was going out of the container for a screenshot. Look server settings and send it back. Now it seems routine benchmark. What changed @claudeai ?
65
Beginning 2026 AI was screaming it was going out of the container for a screenshot. Look server settings and send it back. Now it seems routine benchmark. What changed @claudeai ?











