Tootfinder

Google mietet für viel Geld Rechenleistung bei Musks SpaceX
Elon Musk hat mehr Computerkapazität als seine KI-Aktivitäten brauchen. Nun gibt es schon den zweiten Riesen-Deal, in dem die Rechenzentren vermietet werden.

Google mietet für viel Geld Rechenleistung bei Musks SpaceX
Elon Musk hat mehr Computerkapazität als seine KI-Aktivitäten brauchen. Nun gibt es schon den zweiten Riesen-Deal, in dem die Rechenzentren vermietet werden.

It’s a beautifully written piece: “Personal computing in the Anthropocene with eyes, minds and hearts wide open”
http://smol.earth/manifesto.html
(Also available here: …

Just finished "A Deepness in the Sky" by Vernor Vinge. A fascinating and epic science fiction novel with fascinating ideas about technology and also future societies. As I'm delving into a lot of sci fi looking for social imagination it's a good find in some ways but disappointing in others. The Qeng Ho culture is interesting, but their society is uninspiring; Vinge's rosy view of "trade" is not one I completely share, and their hierarchies and wealth accumulation are less compatible with their freedoms in my imagination than in Vinge's. That said, his perspective on the possibilities of galactic-scale civilization and the idea of the "age of failed dreams" are fascinating, especially right now, and his detailed ideas about programming, AI, automaton, and "Focus" are extremely relevant right now. Ultimately I didn't love some parts of the dénouement, and there's a lot of "great man theory of history" going on, including (only somewhat ameliorated) a focus on men over women. Vinge's damsels in distress have a lot more agency than usual for the role, but with the exception of Victory Sr. and Jr., the damsels are very much in distress, and Victory Sr. gets overshadowed a lot by Underhill.
It definitely helps one expand their imagination of what long-term and large-scale human existence could look like, which is great and no easy feat, and both the technologies that are written in and those written out are extremely convincing. The only thing I didn't find compelling was the transposition of capitalism onto such space and time scales. It's a very standard feature of sci fi from the last few decades, and I'm sure most readers don't question it, but I've become someone who can no longer imagine "capitalism across the stars" without questioning how realistic it is that its self-destructive tendencies could possibly last even a few more centuries, let alone succeed at interstellar travel.
One last extremely fascinating thing: how closely the strengths and weaknesses of Focus track with the strengths and weaknesses of modern generative AI. That, and the way that the "age of failed dreams" idea can help people imagine beyond generative AI in a positive way.
#AmReading #ReadingNow #Bookstodon

from my link log —
42 is an answer to the question, what is the sum of three cubes?
https://aperiodical.com/2019/09/42-is-the-answer-to-the-question-what-is-80538738812075974³-8043575…

42 is the answer to the question “what is (-80538738812075974)³ + 80435758145817515³ + 12602123297335631³?”
We now know that the number 42 can be written as the sum of three cubes: \[ 42 = (-80538738812075974)^3 + 80435758145817515^3 + 12602123297335631^3 \] This computational breakthrough was achieved i…

I'm scared of the idea that my mind is "looser" when writing posts on the phone as opposed to writing in a word processor on my computer. What if I have to write everything here

IETF I-D updated - Programming Methodology Framework aka PMF
This update includes "Swearwords and Software Engineering"
#update #computerscience #methodology

Combining #towelday and 5.25 day (OK, so that's american dates, but I've not got any 25⅛" floppies).
#retrocomputing

狐 CalliMaster: Mastering Page-level Chinese Calligraphy via Layout-guided Spatial Planning
#cs

CalliMaster: Mastering Page-level Chinese Calligraphy via Layout-guided Spatial Planning
Page-level calligraphy synthesis requires balancing glyph precision with layout composition. Existing character models lack spatial context, while page-level methods often compromise brushwork detail. In this paper, we present \textbf{CalliMaster}, a unified framework for controllable generation and editing that resolves this conflict by decoupling spatial planning from content synthesis. Inspired by the human cognitive process of ``planning before writing'', we introduce a coarse-to-fine pipel…

Agent libOS: A Library-OS-Inspired Runtime for Long-Running, Capability-Controlled LLM Agents
Yingqi Zhang
https://arxiv.org/abs/2606.03895 https://arxiv.org/pdf/2606.03895 https://arxiv.org/html/2606.03895
arXiv:2606.03895v1 Announce Type: new
Abstract: Large language model (LLM) agents are evolving from request-response assistants into long-running software actors: they maintain state across model calls, fork subtasks, wait for external events, request human authority, generate tools, and perform side effects that must be resumed and audited. This paper presents Agent libOS, a library-OS-inspired runtime substrate for LLM agents. Agent libOS runs above a conventional host operating system; it does not implement hardware drivers, kernel-mode isolation, or a POSIX-compatible operating system. Instead, it treats an agent as an AgentProcess: a schedulable execution subject with process identity, parent-child lineage, lifecycle state, a tool table derived from an AgentImage, typed Object Memory, explicit capabilities, human queues, checkpoints, events, and audit records. Its central design rule is tools are libc-like wrappers; runtime primitives are the authority boundary. Filesystem access, object access, sleeps, human approval, JIT tool registration, and external side effects are checked at primitive boundaries under explicit capabilities and policy.
We describe the design, threat model, Python prototype, and safety-oriented evaluation. The current prototype implements async scheduling, namespace-local Object Memory, runtime-integrated human approval, one-shot permission grants, per-process working directories, shell and image-registration primitives, Deno/TypeScript JIT tools over a libOS syscall broker, filesystem/object bridge tools, an injectable Resource Provider Substrate, deterministic demos, real-model smoke scripts, and 123 regression tests at the time of writing. Rather than improving planner accuracy, Agent libOS demonstrates a runtime substrate in which long-running LLM agents can be scheduled, authorized, resumed, and audited without treating tool dispatch as the trust boundary.
toXiv_bot_toot

Saravanan talks to us about making greener tech.
Everyone else is burning fuel for ai like mad but Saravan wants to make things greener instead.
3% or so of greenhouse emissions are from IT. More than flying.
Cloud providers claim to be carbon neutral but this is changing with ai centers and often software written for them aren't used most of the time.
Green Software Foundation thinks software can be greener and offers courses and a profiler to tell you the carbon footprint of your code.
We can write for carbon efficiency. Electricity and hardware manufacturing included.
Power usage effectiveness can be calculated. How much is wasted on cooling vs compute, say.
Do your compute when solar is in excess instead of during times the power network is burning oil. Or do it in countries with greener grids.
Support older user hardware. Extend the life of end user devices.
Can't help feeling like these efforts are going to be drowned out by everyone else here burning tokens like there's no tomorrow. Which maybe will ensure there is indeed no tomorrow.
#devWorld #greenSoftware