Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
mbil
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
10 ms
·
91.
▲
by
mbil
3y ago
I imagine it's not long before we see hyper-targeted commercials where the actors look like us, live in our city, etc.
92.
▲
by
mbil
3y ago
Is it $50 for the seed packet or the actual plant?
93.
▲
by
mbil
3y ago
A little off topic, but the 2019 movie Sound of Metal explores some of deaf culture and is a pretty excellent film for anybody interested.
94.
▲
by
mbil
3y ago
I have a RadCity 4 with fenders and a basket. It’s much heavier than a regular bike. But it’s quite fun to ride and can hit 25mph. My city is rather flat and has passable bike infrastructure. I haven’t used it all that much simply because I
95.
▲
Candidate for MontCo Court of Common Pleas Uses Some Unfortunate AI
(old.reddit.com)
2 points
by
mbil
3y ago
|
0 comments
96.
▲
by
mbil
3y ago
Yeah the project name is kinda clickbaity, ironically. Love the idea though
97.
▲
by
mbil
3y ago
I imagined the AI would be reading the steps aloud to you, and so would be aware of your progress. I don’t think an AI assistant precludes the recipe being projected tho, just as in the driving example it wouldn’t replace an on screen map.
98.
▲
by
mbil
3y ago
I agree. I think apps that would initially benefit from LLM-powered conversational interfaces are those that have the following traits: - constrained context - part of a hands-free workflow A couple use-cases I have been pondering
99.
▲
by
mbil
3y ago
Do you mean use the LLM as a post-processing step within a ChatGPT conversation? Or generally (like as part of Whisper)? If it’s the former, I’ve found that ChatGPT is good at working around transcription errors. Regarding the latter, I agr
100.
▲
by
mbil
3y ago
Interesting that removing sulphur emissions may have resulted in more solar radiation being absorbed by the oceans. Wonder if cargo ships could purposely emit some benign material that would reflect away sunlight. Although I suspect there c
101.
▲
by
mbil
3y ago
> The general rule is one car length for every 10 mph I really wish other cars on the highway observed this rule, but it’s common to see people doing ~80 mph with like two car lengths of buffer. When I myself maintain proper following di
102.
▲
by
mbil
4y ago
This is true today, but will it be true in the future where language models are a common interface between human language (probably voice) and software actions? When I go to my bank's website, it's easy for me to find my tax forms
103.
▲
by
mbil
4y ago
This is awesome. How much effort does it take to go from this to a generalist robot: “Go to the kitchen and get me a beer. If there isn’t any I’ll take a seltzer”. It seems like the pieces are there: ability to “reason” that kitchen is a ro
104.
▲
by
mbil
4y ago
If so hopefully it’s not in SVB
105.
▲
by
mbil
4y ago
this piece reminded me of the way i sometimes regarded the world, maybe specifically the natrual world, as a child; that is, i felt that every piece of it was just bursting with meaning for its uniqueness. e.g. a tiny sun-speckled piece of
106.
▲
by
mbil
4y ago
I'll add that there's no evidence that brains do backpropagation
107.
▲
by
mbil
4y ago
Also lambdaprompt https://github.com/approximatelabs/lambdaprompt
108.
▲
by
mbil
4y ago
Yes, and there are often YC company job postings on the front page, which are technically ads.
109.
▲
by
mbil
4y ago
OpenAI already does charge access to its API. ChatGPT is currently free, but LLM model access has been available through a paid API since before GPT-3.
110.
▲
NYC Slice
(elkue.com)
670 points
by
mbil
4y ago
|
509 comments
111.
▲
by
mbil
4y ago
Not ChatGPT, but this article[0] (which was on HN earlier this year[1]) goes through building a small game in JS using OpenAI Codex. [0]: https://www.shawnmatthewcrawford.com/balloons-the-balloon-cl... [1]: https:/&#x
112.
▲
by
mbil
4y ago
Supervillains? Toddlers? These are nonviolent protestors attempting to draw attention to the imminent, massive threat to our planet’s biological systems. If you find their methods annoying or childish: (a) good, at least we’re now talking a
113.
▲
by
mbil
4y ago
Reminds me of a paper that I saw the other day: "On the Paradox of Learning to Reason from Data"[0] [0]: http://starai.cs.ucla.edu/papers/ZhangArxiv22.pdf
114.
▲
by
mbil
4y ago
My read of the ad, as a former Flexport employee, was that they're looking for engineers who are determined and unafraid to work outside of their comfort zone. Understanding the complexities of international customs and logistics is as
115.
▲
by
mbil
4y ago
I fought it which basically entailed me going to some city government building and showing the guy I was out of town for some of the days and so wasn’t able to keep an eye on my bike. He essentially told me to pound sand and so I ended up p
116.
▲
by
mbil
4y ago
Ah yes, reminded me of when my motorcycle was parked near 7th and South St. There was construction I guess, and my motorcycle was apparently courtesy lifted to the sidewalk. I returned from vacation to ~$700 of tickets for the bike being on
117.
▲
by
mbil
4y ago
After having toyed with it for a couple more days, I noticed it sometimes returns malformed JSON with extra tokens. E.g. {"name": "foo", "company": "bar"}</code> That's the only
118.
▲
by
mbil
4y ago
I tried the five most recent messages I received and it worked as expected for all of them. I imagine that I'll encounter some that don't work, and at that point I'll try adding more examples to make the model more robust.
119.
▲
by
mbil
4y ago
I haven’t explored it any further yet, but that seems like a useful extension.
120.
▲
by
mbil
4y ago
Thanks for the feedback. I just updated the article. Basically I plug those extracted values into a templated response.
More ›