Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
tmnvdb
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
15 ms
·
121.
▲
by
tmnvdb
2y ago
They don't want a motivation letter to be written by an LLM (because it's specifically about the personal motivation of the human candidate) - as far as I can see this not reflect either positively or negatively on their level of
122.
▲
by
tmnvdb
2y ago
LLMs often don't do well on tasks that require composition into smaller subtasks. In this case there is a chain of relations that depend on the previous result.
123.
▲
by
tmnvdb
2y ago
Yes, they show benchmarks in the article linked here. Did you not read it?
124.
▲
by
tmnvdb
2y ago
I would consider unsupervised tool usage an achievement
125.
▲
by
tmnvdb
2y ago
HN is full of people who want to feel smart by complaining.
126.
▲
by
tmnvdb
2y ago
You're not wrong but it feels like bikeshedding at this point.
127.
▲
by
tmnvdb
2y ago
This is a hard question for language models since it targets one of their known weaknesses.
128.
▲
by
tmnvdb
2y ago
> At least there was some understanding behind the scenes of where the numbers came from, and sources would/could be provided. Oh Sweet summer child.
129.
▲
by
tmnvdb
2y ago
You can synthesize the parts to get the whole. Both uses are correct AFAIK
130.
▲
by
tmnvdb
2y ago
Peak HN comment
131.
▲
by
tmnvdb
2y ago
How is using public information "exploitation"? A human researcher with Google would do the same.
132.
▲
by
tmnvdb
2y ago
Eating popcorn while the scaling doubters scramble to move the goalposts for the nth time.
133.
▲
by
tmnvdb
2y ago
Have you even watched the video? This is a new capability and not a trivial one.
134.
▲
by
tmnvdb
2y ago
Do you have any benchmarks to back up your 'feelings'?
135.
▲
by
tmnvdb
2y ago
Only if you are asking questions at the level of a cutting edge benchmark
136.
▲
by
tmnvdb
2y ago
That is the biggest problem on this website, people want to feel smart by 'debunking' things they don't really understand. It leads to a lot of contrarian views with poor signal to noise ratio, especially when the topic is sl
137.
▲
by
tmnvdb
2y ago
The US has different presidents all the time
138.
▲
by
tmnvdb
2y ago
Probably they are doing so well because there are not endless meetings on customer friendly names
139.
▲
by
tmnvdb
2y ago
Those models are not comparable
140.
▲
by
tmnvdb
2y ago
It is not very good though.
141.
▲
by
tmnvdb
2y ago
This is such a weird stereotype. The US was declared in 1776. Obviously it has generational wealth. Unless you think everybody lives to be 300 years old.
142.
▲
by
tmnvdb
2y ago
With all due respect, your idea of Europe seems to be about a 100 years out of date and based on period dramas.
143.
▲
by
tmnvdb
2y ago
This is a bit of a myth, most countries in Europe have better social mobility then the US, where people of different social classes are quite segregated. (Though as always, there is a lot of variation between countries.)
144.
▲
by
tmnvdb
2y ago
The country with by far the highest number of people riding a bike (The Netherlands) does not have any helmet rules.
145.
▲
by
tmnvdb
2y ago
That is partly true, though not equally so in every country (Denmark is a positive example - as usual). Note that this kind of regulation is done on the national level and isn't strictly related to EU membership. On the other hand, sal
146.
▲
by
tmnvdb
2y ago
Regulation is only a small part of the story. Keep in mind that GDPR was implemented in 2018, long after Europe lost the tech battle. The lack of a unified (financial) service sector and hence a very weak and fragmented private capital mark
147.
▲
by
tmnvdb
2y ago
Did you try asking deepseek about June 4th, 1989? Edit: it seems that basically the whole month of July 1989 is blocked. Any other massacres and genocides the model is happy to discuss.
148.
▲
by
tmnvdb
2y ago
I disagree with several of these, and the whole thing does not seem very well-conceived, based on an "incorrect" notion of what AI "should" do to be considered impressive. > LLMs only reliably know what you just told
149.
▲
by
tmnvdb
2y ago
Poland is aiming for 4.7% in 2025 and is supporting Trumps proposal for 5% for NATO. Meanwhile Italy is allowed to buy chips with 1.5% defense spending. It doesn't make sense to think that defense spending is the criterion used to sele
150.
▲
by
tmnvdb
2y ago
You can just put the datacentre in Germany or the Netherlands and use it in Poland.
More ›