mr_toad

6 ms·

61.
by mr_toad 2mo ago
“But they can’t match Intels performance” used to be the rejoinder of the Intel crowd. Now we’ve come full circle.
62.
by mr_toad 2mo ago
It’s not like stock photos of infeasibly attractive people eating ludicrously overstuffed burgers were ever interesting or believable.
63.
by mr_toad 2mo ago
The US bombed North Vietnam into dust, but because they didn’t occupy it with ground forces they simply rebuilt and kept coming. You can’t win a war with just air power.
64.
by mr_toad 2mo ago
> What the hell is the point of human interaction if the "human" just forwards a chat bot's response Forwarding an AI response (or any canned response) sends the message that they don’t want to interact with you. That cou
65.
by mr_toad 2mo ago
Mods might be amused to be described as fairly godmothers.
66.
by mr_toad 2mo ago
Glid vs glided, slid vs slided are examples of the inconsistency of the English language. Glided is more common than glid, and slid is more common than slided. Linguistics says that that makes slide an ‘irregular’ verb, but that’s basical
67.
by mr_toad 2mo ago
If you constrained the inputs of Apple or Nvidia then their sales and profits would tumble (just look at what happened to Apples stock price in the last week because of the memory chip shortage). Off the top of my head I can’t think of anot
68.
by mr_toad 2mo ago
It’s an odd kind of bubble that slowly deflates over the course of a year.
69.
by mr_toad 2mo ago
> Quantum experiments you're always looking at some medium that detects or measures or whatever the original interaction. So you're only interrogating the detector, not the original particles or interaction. Exactly, waveform c
70.
by mr_toad 2mo ago
Observations keep poking holes in those hypotheses (I’m not sure they even count as theories). We keep finding black holes with anomalous masses, or black holes that formed too early. Part of the problem is that we have very little observa
71.
by mr_toad 2mo ago
Headline is unnecessary clickbait, the mean was still signaling that something was wrong.
72.
by mr_toad 2mo ago
> What's there, that politicians don't know already? They need to be ‘reminded’ of the issues that ‘campaign contributions’ are contingent upon.
73.
by mr_toad 2mo ago
Have you seen Microsoft’s price over the last year. If AI doesn’t pan out they’re going to be the next IBM.
74.
by mr_toad 2mo ago
Installed by default is their moat. Trying to arrange a video conference on other platforms means making sure that everyone has the client installed, which can be a real headache. Teams is the common denominator.
75.
by mr_toad 2mo ago
> But you don´t get to to that. You can, and for tax purposes you must. The bean counters hate that you can’t deduct capex.
76.
by mr_toad 2mo ago
We’re gonna need a bigger wall.
77.
by mr_toad 2mo ago
“ It has not occurred to you that your ancestors were survivors and that the survival itself sometimes involved savage decisions, a kind of wanton brutality which civilized humankind works very hard to suppress. What price will you pay for
78.
by mr_toad 2mo ago
https://en.wikipedia.org/wiki/ArianeGroup
79.
by mr_toad 2mo ago
> Non-exclusive use of farmland or wasteland (or even - at least until maintenance complexity is considered - offshore) isn't a major barrier Not now, but with solar power output doubling every few years how long before we run out o
80.
by mr_toad 3mo ago
> The reason they couldn't because they didn't have the metallurgy and the machining to create pistons, valves, etc. It was a turbine. You don’t use pistons in a turbine. They lacked knowledge of turbine blades, the metallurg
81.
by mr_toad 3mo ago
It’s not just government money. Factory workers wages in China are still a lot lower than western countries, and they have a lot fewer labor rights. Also, the government hinders as often as it helps, since they can always take over an indu
82.
by mr_toad 3mo ago
> On this benchmark, a pure LLM generated an accuracy score of zero. Adding RAG, prompt engineering, and agentic AI raised accuracy to the 10+% range. That's awful. The AI in Databricks is much better than that.
83.
by mr_toad 3mo ago
The rush hours are the worst. You're lucky if you hear two songs in 30 minutes it's so crammed full of ads.
84.
by mr_toad 3mo ago
> We can then add in an existing dataset of _known_ incidents, indexed by the same common features, as a training/validation set. Beware of the Anna Karenina principle. Well behaved data might be explicable by the same common featu
85.
by mr_toad 3mo ago
In France castles are all called Châteaux.
86.
by mr_toad 3mo ago
New Zealand was colonised long after castles had become obsolete. There are a few remains of WWII coastal fortifications here and there, but they’re now obsolete too. The Māori built thousands of hilltop fortifications with wooden palisade
87.
by mr_toad 3mo ago
It also lists the Bastille, which was completely demolished and is now a public square.
88.
by mr_toad 3mo ago
It’s missing the extensive Napoleonic-era Western Fortifications on the opposite side of Dover from the much more famous Dover Castle. https://en.wikipedia.org/wiki/Dover_Western_Heights But to be fair they’re much les
89.
by mr_toad 3mo ago
> since records began I guess Possibly since 1086 https://en.wikipedia.org/wiki/Domesday_Book
90.
by mr_toad 3mo ago
https://content.imageresizer.com/images/memes/Family-Guy-Ski...

More ›