Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
bertil
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
31 ms
·
361.
▲
by
bertil
3y ago
I’m assuming, given the circumstances, that you can ask all the investors if they are willing to help.
362.
▲
by
bertil
3y ago
AI market is hot. What makes you say that: no one has that kind of cash?
363.
▲
by
bertil
3y ago
I think that’s the most likely outcome: look for a buyer willing to take all the shares at market value, but also willing to offer shares to creditors willing to keep some Anthropic at a private valuation. Not sure if there’s a risk they’ll
364.
▲
by
bertil
3y ago
Most comments are about using higher-level AI. It’s interesting because I have so much trouble getting Grammarly to work (I thought that announcement would justify the problem), but it’s far more prosaic than that: - the little indicator hi
365.
▲
by
bertil
3y ago
I personally think opinionated high-value data would be more helpful, a bit like classifying US voters into seven groups has proven transformational for electoral campaigns. I’ve seen B2B sales desperately trying to sling their solutions to
366.
▲
by
bertil
3y ago
The top of the free listing is repackaged open-source datasets (economic data, Covid). The top of the paid listing is the usual suspects: weather, geographic (GeoIP and matching elements), and you, CyberSyn, but nothing really exotic yet.
367.
▲
by
bertil
3y ago
I’ve seen a lot of people promising to sell (aggregated, cleaned, private, inferred) data on Snowflake — using the database as a platform. Snowflake themselves have been listing this idea as a key value proposition. This is the first that I
368.
▲
by
bertil
3y ago
I was more surprised not to see any AI-specific benchmarks—sure, the most popular open-source models are from avowed competitors, but there should have been a way to define a re-training task that would be relevant for the wave of ML progra
369.
▲
by
bertil
3y ago
I do think that’s a regulatory play, but as you say — expecting few people will take it, but giving the option will shut down criticism. Some here are worried that their data will still be used to train (ads) models: you can easily remove a
370.
▲
by
bertil
3y ago
The most insightful piece of analysis I’ve ever seen was at Facebook, from the rockstars who were looking at Instagram for the first time — a couple of years after the acquisition, when it was time to apply some informed strategic thinking.
371.
▲
by
bertil
3y ago
You are meant to give priority to the vehicles in the roundabouts. A stop sign forces a two-second pause over yielding. I’m not sure that’s a big hindrance, and given how drivers barely mark stops and ignore yield signs, that might be the o
372.
▲
by
bertil
3y ago
Useful lift, so people, luggage, and fuel.
373.
▲
by
bertil
3y ago
There is an endless supply of that, but it's not much lighter than cold air, so it doesn't have a great lifting power.
374.
▲
by
bertil
3y ago
If anything, he would have to be considered behind Waymo.
375.
▲
by
bertil
3y ago
Helium is expensive, which is why they are considering Hydrogen. Yes, that’s what was in the Hindenburg, and yes, it can go big boom, but that’s the idea.
376.
▲
by
bertil
3y ago
It’s a reasonable idea in practice, but they can’t carry that much weight, so either you land often for showers or water would have to be recycled. They can have a lot of power by having the top covered in flexible solar panels (ignoring th
377.
▲
by
bertil
3y ago
Most of the current ones can carry about a ton and a half, but larger ones would be able to carry a lot more: the Hindenburg could carry 232 t, for instance. The largest blades nowadays are 12 t; the nacelle with the transformer is the expe
378.
▲
by
bertil
3y ago
Land transport is the limiting part: you have to limit size to bridge heights and road width standards that are driven by how wide horse arses were in Roman times. Loading them on boats is expensive because they are non-standard size, and l
379.
▲
by
bertil
3y ago
The most meaningful application for airships is to carry blades, nacelles, and towers for wind turbines off-shore. Direct transport from the factory to the farm would greatly simplify installation. Luxury in an airship is quite relative: we
380.
▲
by
bertil
3y ago
No, but I have you seen the Beijing metro: https://www.youtube.com/results?search_query=Beijing+metro and it sounds like it got a lot better than car transport during the same period.
381.
▲
by
bertil
3y ago
There’s progress because there’s investment. If dozens of billions were poured into public transport, it would improve dramatically. How can I be so so sure? Because that’s what a lot of countries have done, China first among many, and thei
382.
▲
by
bertil
3y ago
Connected software, with very few exceptions, doesn’t attract a very large audience. For rarer situations, you aren’t going to find that person easily. Having worldwide interest groups (like subreddits, like thematic Discord channels) helps
383.
▲
by
bertil
3y ago
The post isn’t about Instagram at the exclusion of the fashion ecosystem: it’s an argument that truncates research that does find that posts that the drop in self-esteem is directly related to seeing posts either from fashion professionals
384.
▲
by
bertil
3y ago
> Facebook's own research has pointed at the harmful effects of Instagram on the mental health of children Facebook’s own research has found both positive and negative effects, depending on the child’s status among their peers. The
385.
▲
by
bertil
3y ago
I’d say: if you have the option, go with Afroman… Better beats.
386.
▲
by
bertil
3y ago
Moderation, yes definitely. There’s are a couple billion users, they post a handful of comments per day and rarely post anything long enough to justify applying a transformer, so “everything posted to Facebook/Instagram” is several ord
387.
▲
by
bertil
3y ago
It’s a story in three steps: 1. The problem is that there is a staggering amount of “things” that can be either sold on Facebook/Instagram/WhatsApp or advertised for (it’s the same service): every film on Netflix, every item for s
388.
▲
by
bertil
3y ago
The author might be speculating, but I worked on that team and his diagnostic is very likely exact.
389.
▲
by
bertil
3y ago
Not guessing, I used to work on that team: it’s based on keywords alone. The list of items is too big and changes too often to handle anything beyond that — more than a trillion edits per month. I wouldn’t be surprised if you could sell end
390.
▲
by
bertil
3y ago
Those mistakes are systematic. I worked on the team that dealt with that, and you pretty much can’t mention exotic animals by name if you sell books, movies, fan art, etc.
More ›