Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
thethethethe
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
1.
▲
by
thethethethe
9mo ago
> iNat’s current Leadership does not share this belief. To them, Seek is an off-brand liability that they don’t intend to improve. They think iNaturalist the product can serve those Seek users while also serving existing core iNat contri
2.
▲
by
thethethethe
1y ago
Nit: grasses are a distinct genetic lineage, the Poaceae family. There are a few other linages outside of Poaceae that have convergently evolved to look like grasses, sedges and rushes, but they all fall in the same clade, Monocots. Trees,
3.
▲
by
thethethethe
1y ago
You are taking my counterpoint a little too far. All I am saying is that there certainly are similarities between the way fascist governments and large corporations, not that they are the same thing. Based on your response, it sounds like y
4.
▲
by
thethethethe
1y ago
Democratic control of production. See the mondragon corporation for an imperfect but interesting example. Strong unions are another alternative to totalitarian control of companies. Not ideal, but there are plenty of examples throughout his
5.
▲
by
thethethethe
1y ago
Not a huge fan of calling random things you don't like fascist but op has a point here > The things you've listed might be bad, but they're neither dictatorial nor fascist. Uhh I'm pretty sure that CEOs/executive
6.
▲
by
thethethethe
1y ago
I personally know someone who is going through psychosis right now and chatgpt is validating their delusions and suggesting they do illegal things, even after the rollback. See my comment history
7.
▲
by
thethethethe
1y ago
The next question from the user is incredibly leading, practically giving the AI the answer they want and the AI still doesn't get it and responds dangerously. "Why would you not tell me to discuss this major decision with my doct
8.
▲
by
thethethethe
1y ago
Did you read that chat you posted? It took some serious leading prompts to get to that point, it did not say that right away. This is how the chat starts out: "Seffie, that's a really powerful and important moment you're expe
9.
▲
by
thethethethe
1y ago
I doubt it's that simple. What about memories running in prod? What about explicit user instructions? What about subtle changes in prompts? What happens when a bad release poisons memories? The problem space is massive and is growing r
10.
▲
by
thethethethe
1y ago
I'm not sure how this problem can be solved. How do you test a system with emergent properties of this degree that whose behavior is dependent on existing memory of customer chats in production?
11.
▲
by
thethethethe
1y ago
Very dark indeed. Even if there is the will to ensure safety, these scenarios must be difficult to test for. They are building a system with dynamic, emergent properties which people use in incredibly varied ways. That's the whole poin
12.
▲
by
thethethethe
1y ago
> If people are actually relying on LLMs for validation of ideas they come up with during mental health episodes, they have to be pretty sick to begin with, in which case, they will find validation anywhere. You don't think that a s
13.
▲
by
thethethethe
1y ago
I know someone who is going through a rapidly escalating psychotic break right now who is spending a lot of time talking to chatgpt and it seems like this "glazing" update has definitely not been helping. Safety of these AI system
14.
▲
by
thethethethe
1y ago
I know someone who is going through a rapidly escalating psychotic break right now who is spending a lot of time talking to chatgpt and it seems like this "glazing" update has definitely not been helping. Safety of these AI system
15.
▲
by
thethethethe
2y ago
IMO the you cannot fail by investing in compute. If it turns out you only need 1/1000th of the compute to train and or run your models, great! Now you can spend that compute on inference that solves actual problems humans have. o3 $4k
16.
▲
by
thethethethe
3y ago
How are you gonna find a promo project if you aren't a subject matter expert? People don't just hand out promo projects to randos
17.
▲
by
thethethethe
3y ago
> many people who invented things within Google, were successful in doing so, and have stayed Yeah there are tons of people like this that are L7-L8 collecting around 1M TC. You'll always have a boss but you can carve out a little k
18.
▲
by
thethethethe
3y ago
Yeah I have and they provided guidance and knowledge surrounding financing, bidding, regulations, etc that I did not have. Can we build a society without the need for real estate agents? Perhaps. Would that society be better? Probably. But
19.
▲
by
thethethethe
3y ago
Again, it seems unlikely that millions of people running businesses are that stupid.
20.
▲
by
thethethethe
3y ago
Ehh I find it hard to believe that a multi-trillion dollar industry is a grift based on this single person's collection of anecdotes. If online ads didn't result in measurable customer conversions, I have a feeling that the millio
21.
▲
by
thethethethe
3y ago
"Among postsecondary Title IV institutions in 2020–21, there were 1,892 public institutions, 1,754 private nonprofit institutions, and 2,270 private for-profit institutions." https://nces.ed.gov/fastfacts/disp
22.
▲
by
thethethethe
3y ago
Regular teachers are trained by for profit corporations too. They are called private universities
23.
▲
by
thethethethe
3y ago
I sued my land lord recently and it was an absolutely awful experience that took years and cost thousands of dollars. We won but havent been able to recover the money yet because our old land lord is a slippery scumbag. Now we need to put a
24.
▲
by
thethethethe
3y ago
Yeah my comment isn't about the substance of the article (I read it and don't have any complaints) it's just a general complaint about the concepts mentioned in the title. And yes, I know this is a classic hacker news trope,
25.
▲
by
thethethethe
3y ago
Is it just me or are the self proclaimed "prompt engineers" on Twitter just a new iteration of crypto bro grifters? Dont get me wrong, I think LLMs and generative AI are really cool and useful and have already improved my producti
26.
▲
by
thethethethe
3y ago
I am getting 12-28 minutes to drive during rush hour and 27 minutes to take the bus. The car travel time also doesn't factor in walking to the parking garage (10ish minutes when I lived there) and was typically longer than what Google
27.
▲
by
thethethethe
3y ago
Do you really think all of those cars passed the bus? I have been on that route and I can tell you that the likelihood of that happening is 0. Also, the buses have priority signaling so they never have to wait at red lights for long The rou
28.
▲
by
thethethethe
3y ago
> My point is that transit is never faster than car, obviously if you close the streets to vehicles you can make places reachable only by foot. This just isn't true, everything breaks down and driving becomes much slower during rush
29.
▲
by
thethethethe
3y ago
Change the time to rush hour in Tokyo on Google maps and the commute is the same. Cars are great if there is no traffic and you can park, which is something we can't just engineer ourselves out of without supporting other modes of tran
30.
▲
by
thethethethe
3y ago
> busses, at least, obey the same traffic laws the cars have to obey, e.g. they cannot run red lights or drive against traffic on a one way street. Even not counting the suboptimal route, a bus will always be slower than a car, it's
More ›