bigfudge

16 ms·

271.
by bigfudge 2y ago
For context for non EU people: > As the commission is the executive branch, candidates are chosen individually by the 27 national governments. Within the EU, the legitimacy of the commission is mainly drawn from the vote of approval that
272.
by bigfudge 2y ago
The median EU citizen is not getting poorer relative to the median US citizen. Don't confuse growing income inequality with "national wealth".
273.
by bigfudge 2y ago
Chromes replacement of Firefox had more to do with aggressive marketing and deliberate poor performance of their web properties like gmail on non-Google browsers. Not sure how much consumer self direction was involved.
274.
by bigfudge 2y ago
Most belt drives are run dry so the oil doesn’t attract dirt in the same way
275.
by bigfudge 2y ago
Also many medium sized companies which are productive and competitive but not public, so never grow to huge sizes. I view this as a feature not a bug… smaller companies have a more direct connection with their workforce and tend to behave b
276.
by bigfudge 2y ago
And yet standards of living in Europe are comparable to those in the US, and preferable at the median. Our attention is captured by speculative valuations of unicorns, and yet people actually need real stuff made, drugs developed and made e
277.
by bigfudge 2y ago
Give the uk a chance. It will take time to unwind some of the more stupid brexit decisions without obviously losing face. There is appetite for it, but it’s probably going to take at least another govt to be able to make consequential decis
278.
by bigfudge 2y ago
Have there never been cases where law enforcement requested that sort of information and used it? I don’t think we know.
279.
by bigfudge 2y ago
In my experience self validation performance and confidence ratings are both poor. I think the problem is that these sorts of formats just aren’t that common in the training data. What does help is to as a series of structured questions per
280.
by bigfudge 2y ago
And yet Italian diabetes rates (and life expectancy) are higher than the US.
281.
by bigfudge 2y ago
In my experience, mostly just buying nice beans and not doing anything stupid with them. The hangover from Italian style coffee is quite strong and means we get a lot of over-roasted beans (or at least, cheap beans over-roasted to disguise
282.
by bigfudge 2y ago
The point of bullshit charges is often to obscure total cost upfront. We are absolutely rotten with them in the UK. Not sure the US is better but I think they are really anticompetitive because they create a cost to discovering the true cos
283.
by bigfudge 2y ago
Because once the current safety scrutiny has passed you might get more trips done by setting the ai to be more aggressive in traffic. Then you are into VW style software updates with a profit motive and no mechanism to hold them accountable
284.
by bigfudge 2y ago
The other problem is the economics flips over once you also have to own a car. The marginal cost of each trip goes down … but if waymo is good enough for 95% of the time the it might not be necessary any more.
285.
by bigfudge 2y ago
There will be bullshit booking charges etc though. The thrust of their point is largely correct. I really wish people were more sensitive to being stiffed like we are by these sorts of stupid charges … that they arent means there is often n
286.
by bigfudge 2y ago
I think you’re wrong about most of the scenarios on your list. And once the market is mature, I can imagine it would be great to be picked up in a minivan after a days cycling somewhere new and not on a loop route. Americans have become emo
287.
by bigfudge 2y ago
The cheap etymotic plugs are great value. I always have a pair on my keyring. The advantage over foam plugs is that the attenuation is more linear so you don’t feel so weird wearing them.
288.
by bigfudge 2y ago
You definitely can do that. It’s just sometimes simpler to dump lots of stuff in context and then check it wasn’t made up. It like the idea of using markdown footnotes. I think that would word well - ChatGPT does handle markdown really wel
289.
by bigfudge 2y ago
RAG isn’t just search though. I can be using retrieval for generating content, but I think RAISE kind of restricts the term to question answering which is not the e only application.
290.
by bigfudge 2y ago
How? Do you just mean a simple way that they could say I don’t know but tend not to?
291.
by bigfudge 2y ago
HYDE is a related technique. Ask the model to generate a response with no context, then use this for semantic search agains actual data and respond by summarising these documents.
292.
by bigfudge 2y ago
In my experience, it really does reduce the risk of hallucination, especially if paired with a prompt which explicitly instruct the model to use only facts from the context window. Another strategy is to provide unique identifiers for the
293.
by bigfudge 2y ago
I recently started watching fallout and it reminded me of a book I read about a future religious order which was piecing together pre-bomb scientific knowledge. It immediately pointed me to the Canticles of Leibovitz (which is great btw). G
294.
by bigfudge 2y ago
But the plans were on display…” “On display? I eventually had to go down to the cellar to find them.” “That’s the display department.” “With a flashlight.” “Ah, well, the lights had probably gone.” “So had the stairs.” “But look, you found
295.
by bigfudge 2y ago
It’s not obvious how any of those would do anything but better approximate the average of a noisy dataset. RLHF might help, but only if it’s not done by idiots.
296.
by bigfudge 2y ago
I fear you might be true. I do worry though that a big reason LLMs are good and answering tech questions is that they’ve read SO
297.
by bigfudge 2y ago
That was a really long time ago though. It’s seems like it’s been awful for at least 10 years.
298.
by bigfudge 2y ago
And yet social mobility is higher in the EU vs the UK and US.
299.
by bigfudge 2y ago
Can you provide some numbers to back this up. This plot suggests Germany and USA have equivalent rates of economic inactivity (which is the term to describe what you are talking about). Some Southern European countries have higher rates, b
300.
by bigfudge 2y ago
This is misleading. In Denmark at least wages are often set through collective bargaining. Wage inequality and rates of low pay are much lower in Denmark than US. They just achieve it by slightly different means. In this case it is state ma

More ›