Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
victor9000
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
9 ms
·
91.
▲
by
victor9000
3y ago
I can't understand how anyone feels comfortable giving up that type of information.
92.
▲
by
victor9000
3y ago
So many people racing to crown themselves as the trusted censor of generative models.
93.
▲
by
victor9000
3y ago
Where will I go? Anywhere but Flexport
94.
▲
by
victor9000
3y ago
Seems like a great time to start poaching flexport talent. Now that leadership has shown its true colors, I can't imagine anyone working there is feeling very confident about their future.
95.
▲
by
victor9000
3y ago
It sounds like a great way to miss out on talent that can command a salary higher than the average, while simultaneously attracting talent for whom this is a good deal.
96.
▲
by
victor9000
3y ago
By employing high-pressure sales tactics to trigger fomo in candidates, except it never worked. He climbed the ladder at a well known tech company, almost to the very top, and tried to use those same "skills" in a startup context
97.
▲
by
victor9000
3y ago
The failure modes I've encountered: 1. Toxic team members who can't be fired. These people will reduce morale down to zero and force your best people to leave. In one case it was the CEOs long-time friend, and in the other case
98.
▲
by
victor9000
3y ago
I'm thinking of cases where the pdf is accessed over a network. Like integrations with systems that do billing, invoicing, taxes, tickets to a game, rent receipts, pulling pdfs from your email, pulling pdfs from S3, almost everything?
99.
▲
by
victor9000
3y ago
A more thoughtful discussion on the subject: https://opensource.stackexchange.com/questions/1717/why-is-c...
100.
▲
by
victor9000
3y ago
So Free, as in 4 random letters that were strung together.
101.
▲
by
victor9000
3y ago
Oof, unfortunately the Alfredo license kills a lot of use-cases for this project.
102.
▲
by
victor9000
3y ago
This also works when reviewing resumes. If you see enough candidates list xyz as a skill, then it's probably worth looking into.
103.
▲
by
victor9000
3y ago
Yeah, they went from v1 to regulatory capture in the span of months
104.
▲
by
victor9000
3y ago
It definitely feels like the bias is coming from the prompt itself
105.
▲
by
victor9000
3y ago
I played this game and holy cow, it's not just biased but it's consistently biased. Granted we don't know what the prompt looks like, so the biases could be embedded in the prompt as opposed to the model.
106.
▲
by
victor9000
3y ago
My original comment pertains to how employees can execute on a mission, not how a market will respond to what they produce. Meaning while increases in morale can be conducive to higher productivity, someone working in a factory can't
107.
▲
by
victor9000
3y ago
They say privacy, but this sounds like a request to make censorship more censor friendly. Edit: People seem to disagree, but is this not a tool for wiping ideas from your training data? Meaning they want to train on Wikipedia and remove co
108.
▲
by
victor9000
3y ago
It's amazing how you switched metrics from production numbers to sales numbers to make a point. Nevermind the number of unrealized buyers who are passing on the brand because of it's toxic leadership, but perhaps you can switch m
109.
▲
by
victor9000
3y ago
It's amazing what competent people can achieve once they stop being micromanaged by a narcissist.
110.
▲
by
victor9000
3y ago
Bye bye Birdie!
111.
▲
by
victor9000
3y ago
Mainly to distract myself as a way to destress
112.
▲
by
victor9000
3y ago
I assume it's done with fine tuning. Meaning you come up with responses that capture topics you want to avoid, then let the model convert these unwords into their corresponding latent space, and train it to exclude that space altogeth
113.
▲
by
victor9000
3y ago
Documentation is a necessary skill at senior levels in large orgs. There will never be a shortage of interested parties eager to vampire away your cycles. Being able to send them docs, diagrams, etc. will save you so much oxygen.
114.
▲
by
victor9000
3y ago
This is what happens when you put your country up for sale. Money equals speech and corporations are people too, so why wouldn't they buy houses?
115.
▲
by
victor9000
3y ago
Well, how else are they supposed to reach 99% accuracy!?
116.
▲
by
victor9000
3y ago
My favorite are executives whose products are paying customers of generative AI (aka no moat), yet they parrot platitudes as if they were at the driver's seat.
117.
▲
by
victor9000
3y ago
Why not use the image to provide a better "scan me" visual?
118.
▲
by
victor9000
3y ago
Honestly, this could afford a lot of liberty. If no one's really listening then you could treat it like a jam session and polish new work.
119.
▲
by
victor9000
3y ago
Nevermind the arduous conditions they subject warehouse workers to.
120.
▲
by
victor9000
3y ago
lol, pathetic
More ›