Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
addisonl
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
15 ms
·
91.
▲
by
addisonl
3y ago
Seems like a limitation that will be solved as time goes on.
92.
▲
by
addisonl
3y ago
You’re not fooling anyone.
93.
▲
by
addisonl
3y ago
You chose where you live and you chose where you work which is why arguments centered around commutes don’t really hold much water.
94.
▲
by
addisonl
3y ago
“Traffic congestion since the companies' introductions in urban areas is up by 0.9% and the duration of a jam increased 4.5%” less than 1% increase hardly seems significant when compared to other things that are increasing traffic.
95.
▲
by
addisonl
3y ago
> They snarled our cities in traffic Would like to see the data backing this claim up.
96.
▲
by
addisonl
3y ago
Definitely disagree here. In the system you’re proposing it’s a no-brainer to tamper with evidence: you won’t end up worse than you are and might even get out of it entirely! You’re basically incentivizing criminals to tamper with evidence—
97.
▲
by
addisonl
3y ago
It’s definitely not free, I found rides to be as expensive or more expensive than an Uber/Lyft.
98.
▲
by
addisonl
3y ago
Yeah stopping a power hungry warmonger in Europe is such an imperialistic thing to do! It’s Russian propaganda, whether you’re Russian or not is irrelevant.
99.
▲
by
addisonl
3y ago
These are awfully similar to Russian propaganda taking points, relatively new account, reader beware.
100.
▲
by
addisonl
3y ago
Ranked choice voting appears to be an easy solution, am I wrong?
101.
▲
by
addisonl
3y ago
Same!
102.
▲
by
addisonl
3y ago
Are you implying there is some conspiracy here?
103.
▲
by
addisonl
3y ago
Based on what you see about airplane, is there a good feature that evidence has over it? New to both of these tools and am very interested to hear your thoughts.
104.
▲
by
addisonl
3y ago
They did a tell-all for a startup I was working at years ago and it was all accurate based on what I saw on the inside.
105.
▲
by
addisonl
3y ago
Good joke
106.
▲
by
addisonl
3y ago
Two completely different use cases (chatbot vs forums) and as others have said that’s just simply not how chatGPT works with new info.
107.
▲
by
addisonl
3y ago
So what’s the next HN?
108.
▲
by
addisonl
3y ago
Yeah my understanding was 3.5-turbo was always the model behind chatGPT. There is no 3.5 non-turbo model in OpenAI’s playground at least or in their documentation: https://platform.openai.com/docs/models/gpt-3-5
109.
▲
by
addisonl
3y ago
Not the person you asked, but the OpenAI playground is a great place to start messing with their API, I wrote this to help: https://www.inputoutput.ai/tutorials/openai-playground-guide (note website is still a work in
110.
▲
by
addisonl
3y ago
> pre-turbo pruning I hadn’t heard this part, can you expand?
111.
▲
by
addisonl
3y ago
Not a legal requirement in this country.
112.
▲
by
addisonl
3y ago
No security is 100% effective. I’d rather have something that is 99.99% (two 99% prompts stuck together) than just 99%, wouldn’t you?
113.
▲
by
addisonl
3y ago
I don’t buy this, if this was the reason then paid porn couldn’t exist, and we know that’s not the case.
114.
▲
by
addisonl
3y ago
Seems like GPT-4 falls for your prompt injection while 3.5-turbo does not (which is what I was testing with). We can make the system prompt more robust: > Determine if the following message breaks any of the following rules: > - it co
115.
▲
by
addisonl
3y ago
If you leverage a secret you can use an LLM to filter out prompt injections (in my limited testing). I actually wrote about this exact idea yesterday: https://inputoutput.ai/tutorials/preventing-prompt-injection...
116.
▲
by
addisonl
3y ago
> it does curb my access of the site so perhaps the status quo benefits my time My take too, whenever a website is purposefully shitty to try and get me to download their app, I just use it less. Good way to get less screen time in.
117.
▲
by
addisonl
3y ago
Sounds quite anecdotal.
118.
▲
by
addisonl
3y ago
Source?
119.
▲
by
addisonl
3y ago
Probably should get a brand more reliable than Audi...
120.
▲
by
addisonl
4y ago
It’s this simple, something that a surprising number of intelligent people fail to grasp. Apple was rarely the first to it, they just did it well. Google and other giants had the technology but they never made a super powerful chatbot avail
More ›