I have a dozen data scientists who are comfortable with Pandas and requests but have never written code to open a socket, who use Docker almost daily but aren't sure what a certificate is, and who understand that machines have IP addresses but haven't ever been shown how DNS works. They would like to learn some more about computer networks—would someone in/around Toronto be willing to teach a one-day hands-on workshop in person, virtual, or hybrid? If so, please DM me—thx.
They state that customers "contractually consent" to such use, but good luck finding it in their Terms of Service. There also doesn't appear to be a way to withdraw consent, but I may have missed that.
@KFosterMarks I remember a Sun Microsystems rep talking about "programmer comfort" in the late 1980s (more specifically, as something they had that IBM and SGI didn't). I also remember the phrase "engineering ergonomics" from about the same time, but can't recall context. The idea was definitely used by Unix fans as a stick to beat up Windows in that period even if the name varied.
turns out "satchel" isn't as widely used a word as I thought, but so far everyone seems to know what a "knapsack" is, so "knapsack-scale computing" maybe?
Looking for a word to use in contrast to "exascale" to mean "the laptop-sized computing that is all 99.9% of researchers need". Thought about "human-scale" but that doesn't have enough zing. Suggestions?
I find it difficult to get excited about telerobotic brain surgery when an increasing number of people are dying of preventable infections. Similarly, I'm unenthused about discussions of systems thinking applied to digital transformation of enterprise architecture when the overwhelming majority of programmers don't have meaningful test suites. 1/
Irugalbandara et al 2024: "A Trade-off Analysis of Replacing Proprietary LLMs with Open Source SLMs in Production" https://arxiv.org/abs/2312.14972 "We present a systematic evaluation methodology for…Small Language Models (SLMs) and their tradeoffs when replacing a proprietary LLM APIs…across 9 SLMs and 29 variants, we observe competitive quality-of-results for our use case, significant performance consistency improvement, and a cost reduction of 5x-29x when compared to OpenAI GPT-4." #nwit
I program, write, and teach. Co-founder of Software Carpentry and It Will Never Work in Theory; co-editor of The Architecture of Open Source Applications.