laminatedsmore

I didn't get this either until I looked at the linked doc, which breaks it down more obviously imo (section 6.1 for examples).

Looks interesting to me, seems nice to be able to handle refreshing content with pure http as described here rather than with web sockets.

https://datatracker.ietf.org/doc/html/draft-toomim-httpbis-b...

Thanks!

That's interesting, thanks for sharing. It's got me interested in trying this out for myself, do you happen to have / know of any other writing about workflows for setting up remote dev environments using prox mox?

"grateful that literacy, the printing press, computers and the internet became normalised before this notion of "harm" and harm prevention was"

Printing Press -> Reformation -> Thirty Years' War -> Millions Dead

I'm sure that there were lots of different opinions at the time about what kind of harm was introduced by the printing press and what to do about it, and attempts to control information by the Catholic church etc.

The current fad for 'safe' 'AI' is corporate and naive. But there's no simple way to navigate a revolutionary change in the way information is accessed / communicated.

I'd say that knowledge (stored as weights) is acquired through experience (starting from a mostly blank start, weights are formed through exposure to training data). For me that's a useful mental model which helps me think about what a LLM is and isn't.

'Understanding' has an ambiguous meaning, while thought and senses are certainly not applicable.

But it's a semantic discussion about a novel and not well understood topic so :shrug: :)

The googled definition says "the mental action or process of acquiring knowledge and understanding through thought, experience, and the senses.", which I think fits ok.

Do you have a word that you prefer?

I had a chat with GPT about this and it came up with the term 'data grounded cognition' to describe an 'intelligence' that is derived purely from (and expressed through) statistical patterns in data.

I quite like the term, and it seems quite unique (perhaps cribbing from 'grounded cognition' though that's an entirely different idea AFAIK)

If you feel this is a problem for you, and you haven't already, it's worth checking with a doctor for blood tests. I felt similarly in the mornings and being diagnosed + treated for an underactive thyroid was a huge improvement

afaik the aim is to provide enough warning to protect transformers, I guess by powering them down in some way?

Not sure it answers your question but in case interesting here's a link to a UK document describing 'space weather preparedness strategy'

https://assets.publishing.service.gov.uk/government/uploads/...

We try to use a Community of Practice framework, e.g. https://en.m.wikipedia.org/wiki/Community_of_practice. Active members of a CoP might then decide to set up a book club etc.

It's quite easy to draw a line from org principles ("we're always learning" etc) to active participation in a CoP, so helps align incentives on a team level (we can spend time on things outside of our immediate needs) and an individual level (it's an obvious way to get performance review evidence)

Conda has been a lifesaver for me in the past, but it got so slow in ~2019 (minutes+ to resolve dependencies) that I've switched back to pip whenever possible. Maybe things have been resolved now though?

E.g. https://github.com/conda/conda/issues/8087, https://www.anaconda.com/blog/understanding-and-improving-co...

This is cool, but the high cost of using step functions ($0.025 per 1,000 state transitions) makes me pretty wary of using it for anything that doesn't really need the orchestration functionality (i.e. long term wait / retry logic)

Do 4xx responses count against your SLO? For me they don't, but an abnormal increase might still signify that something is actually wrong. (I haven't yet found a useful tool for highlighting this kind of abnormality though)

I also like to chain functions in sequence like this, I use the toolz library to do it in Python.

https://toolz.readthedocs.io/en/latest/api.html#toolz.functo...

"I.e. pipe(data, f, g, h) is equivalent to h(g(f(data)))"

From https://www.ecmwf.int/en/about/media-centre/science-blog/201..., more plain language discussion about low cloud at https://blog.metoffice.gov.uk/2017/04/07/why-is-forecasting-...

"Convection is complex and hard to model because it involves multiple processes, many of which take place at scales smaller than the model’s grid cell, so they need to be parametrized. The assumptions that have to be made when parametrizing convection explain many of the systematic errors in convective precipitation."

(in other words modelling physical interactions on a relatively small scale (1km) is hard)

That link explains that this is just a protocol update not a security thing. (confusing word choice imo)

I think a python implementation of flow is in the toolz library as compose_left. I find myself often using the similar pipe function.

- https://toolz.readthedocs.io/en/latest/api.html#toolz.functo...

- https://toolz.readthedocs.io/en/latest/api.html#toolz.functo...

Thinking about this further, the benefit of these restrictions is in avoiding conflicts where dependant infrastructure is updated in place before the blue green has completed. Lack of inputs and outputs still seems killer though.

The list of restrictions in the user guide seems painful - cannot be used in the same template as nested stacks, cannot start a blue green deploy if other infrastructure would be modified at the same time, cannot import properties from other stacks, cannot export output properties. I was interested, but can't imagine using it with these restrictions.