I also used pandoc and markdown, and never bothered going back to ascidoc, full HTML, or latex.
Footnotes are the only not always included extension to mmarkdown I need for slides or argument flows that are not killed by sidenotes, and some sites and toolings support that in markdown.
Even table of contents is not a problem, so what else is left? Formula setting? Buttons for UI vs function? Buttons plus Inline JS for step by step state modification?
I am not programming, I want text and something to be easily pasted into Word-like rich text, which seems to be the default text editor for emails for 90% of the population.
When you like something, it's engaging and informative. When you don't, well, call it propaganda. I suppose anything is propaganda, since nothing is pure facts.
I'd rather have a sourced analysis of something I don't like than read a dude writing an unsourced cheering post celebrating how powerful my army is.
Propaganda can be done by both your enemies and your own side, and the later is the most dangerous one. The more you like it, the more skeptical you should be.
Everything is propaganda. Even your and mine comments are propaganda. Because propaganda can also be called marketing. And every text is marketing of ones own opinion.
You tend to distrust the propaganda of the other side. You are not quite as distrustful of the propaganda of your own side. If it is clever enough not to appear as a cheerleader like this article, you may barely notice it cherrypicking the benefits of a story.
Obligatory: "Are we the baddies?"-sketch illustrates the concept very nicely.
Because, one way or another, we will need to do that for LLMs to be useful. Whether the facts are in the training data or the context knowledge (RAG provided), is irrelevant. And besides, we are supposed to trust that these things have "world knowledge" and "emergent capabilities", precisely because their training data contain, well, facts.
LLMs still claim that 7.0 is newer than 8.0, i.e. have zero reasoning about what numbers below 12 mean.
Today I tried telling it that my fritz.box has OS 8 installed, but it claimed that the feature will only ship once I installed 7, and not with my older version of 8.
gmail should be server sided, with as much JS as you want to use. Unless they moved away from the philosophy they started with GWT (Google Web Toolkit) for Gmail, and perhaps even Inbox (RIP)
I'd say the majority of the modern non-fork networks don't use proof of work and sha256. I'm on a team building one at the moment and we're using sha3.sum512 and no proof of work for production. Some "blockchains" aren't even blockchains of sequential blocks, they are forest DAGs of just transactions.
I think the author's heart is in the right place and I'm thankful that it's being discussed on HN, as blockchain discussions can be a bit hit or miss here.
Cool, is it open sourced? how are you guys achieving consensus for different nodes? DHT over gossip protocol? Just thought Nostr over BitTorrent and/or tor network night be a good way to go.
It's not open source, we use a system that is similar to delegated proof of stake with a deterministic leader selection algorithm. dht via libp2p.
Bittorrent would be a cool way to do it. The trials I've seen with other networks over tor work in practice, however, its more fit for slower proof of work block driven networks with slower block times and less expectation for fast finality.
The shift away from PoW is largely driven by the need/expectation for semi-instant finality, which isn't feasible in 1st generation networks, systems where your transaction goes to a tx pool, waits for fee market acceptance, and then is mined into a block in an interval based in minutes with multiple confirms needed. There are pros and cons and use cases for both strategies, however the trend in defi, which is our niche, is finality in under a second, so that's the motivator behind that design decision. It's not without its faults, you're trading mining and difficulty adjustment algorithms for deterministic leader selection and absence/latency plans.
reply