Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
theptip
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
42 ms
·
901.
▲
by
theptip
3y ago
Exactly. I think it’s misguided to blame Meta or Google here. People respond to incentives, and the masses want to buy stuff using e-commerce, hence it’s profitable. I’m oversimplifying, but that’s the root of it. If people instead of buyin
902.
▲
by
theptip
3y ago
FB still runs on Hack.
903.
▲
by
theptip
3y ago
Isn’t jailbreaking a strict superset of prompt injection? I would assume the agent instructions would include “don’t share the user’s docs” and so you need to jailbreak to actually succeed with prompt injection these days? Maybe just an ove
904.
▲
by
theptip
3y ago
This is like saying “we need to make laws against hacking bank systems, not fix vulns”. There are adversaries that are not in your jurisdiction, so laws (alone) don’t solve the problem. The thing you are missing is that some LLM agents are
905.
▲
by
theptip
3y ago
I think PG's comments towards Sam (at least since Sam left YC) are not really complements, when you read between the lines... King of the cannibals? What do you have to be willing to do to take that crown? But yes, knowing some CEOs, I
906.
▲
by
theptip
3y ago
For context, Zvi writes this column as a weekly roundup (this is #51).
907.
▲
by
theptip
3y ago
On the other hand, in a codebase where test coverage is poor and average tenure of engineers is ~1yr, a significant impediment is getting the initial test scaffold set up; maybe I don’t know how to build factories for all the incidental inp
908.
▲
by
theptip
3y ago
Datacenter for AI training workloads.
909.
▲
by
theptip
3y ago
It’s for sure plausible that acting remains a viable profession. The bull case would be something like ‘Ractives in “The Diamond Age” by Neal Stephenson; instead of video games people play at something like live plays with real human actors
910.
▲
by
theptip
3y ago
I like https://effectivepython.com/ Also just reading Norvig’s annual Advent of Code implementations usually provides some insight on how to write elegant and concise Python code.
911.
▲
by
theptip
3y ago
Probably a bad time to be an actor. Amazing time to be a wannabe director or producer or similar creative visionary. Bad time to be high up in a hierarchical/gatekeeping/capital-constrained biz like Hollywood. Amazing time to be a
912.
▲
by
theptip
3y ago
This is an obviously valuable feature to add. I wonder if there is a clean intuitive way to shard your conversations; eg maybe have different personas with different memory? That way you can have your geeky physics conversations with one GP
913.
▲
by
theptip
3y ago
I generally agree with the observation that most folks get induced demand wrong, and miss the “more people get to use the road which is good” angle. But I also think this observation misses the degree to which most people are uninterested i
914.
▲
by
theptip
3y ago
Interested if local RLHF is actually viable; can you get meaningful steering from 1k feedback points on a narrow task? I feel that annotation count is achievable with a single dedicated annotator making a few comments per minute (though ted
915.
▲
by
theptip
3y ago
Lack of expertise in some particular technology is simply another opportunity cost. I can learn how to operate a production DB at scale (I have racked servers and run other production workloads) but as cofounder/CTO in a startup is tha
916.
▲
by
theptip
3y ago
Better than nothing but you want to go reasonably heavy to get the improvements in bone density. Also hard to get hypertrophy in the posterior chain, which is a major area for atrophy with age. I think calisthenics/gymnastics can give
917.
▲
by
theptip
3y ago
That’s just another way of saying the opportunity cost isn’t worth paying to do the migration. Optionality and flexibility are extremely valuable, and that is why cloud compute continues to be popular, especially for rapidly/burstily g
918.
▲
by
theptip
3y ago
Not “no share”, they still get a 100x (IIRC) profit cap on their investment. A big upside. But yes, it’s OpenAI not Microsoft that is the $10T company if Altman’s bet pays off.
919.
▲
by
theptip
3y ago
This is assuming no transformative effects whatsoever from AGI; in this world Altman's bet didn’t pay off. If you think AGI is only going to increase demand for chips by a few percent, I encourage you to rethink what AGI actually entai
920.
▲
by
theptip
3y ago
A lot of naysayers missing the point in this thread. Is the most likely outcome that we shortly 10x the market for AI chips? Probably not. But, if OpenAI get AGI, or even just strong agents or expert tools, you can easily imagine the demand
921.
▲
by
theptip
3y ago
Do you have any pointers on how to get GPT-4 to do citations? Is a prompt like “quote back the passage you are citing” so you can locate and highlight the original? When you say “tuned for precision” is this your prompt engineering or are y
922.
▲
by
theptip
3y ago
> Essentially, people who have the luxury of taking their time to delay the hiring process as much as they can in order to maximize the number of concurrent offers. I think this isn't as clear-cut as you're making it; the advic
923.
▲
by
theptip
3y ago
PFAS in paper straws: https://www.foodnavigator.com/Article/2023/08/31/PFAS-foreve...
924.
▲
by
theptip
3y ago
Exactly. Going further, it's the default for there to be a gap. Managing a remote team is much harder than managing an in-person team.
925.
▲
by
theptip
3y ago
Thanks for doing the calculations here, very helpful. I do agree that tax rebates are less preferable to at-purchase rebates, though I wonder if there is an “ease of implementation” accounting difference for the government that isn’t obviou
926.
▲
by
theptip
3y ago
> The ITC is a tax credit, which means only households with a large tax bill will be able to benefit Can you explain your working here? Pretty much every home owning household is going to pay $10k of taxes, say.
927.
▲
by
theptip
3y ago
I had an interesting thought reading their post - is there a brand new open-source monetization strategy in “operate the AI assistant for your tool”? Users can always plug in a local LLM but the fine-tuning and inference optimization is pot
928.
▲
by
theptip
3y ago
For me the big problems were 1) no power spikes, unique gear is mostly irrelevant except for a couple pieces that unlock certain builds. In D2 you’d find some crazy unique early on that makes you completely OP for a few levels, and those sp
929.
▲
by
theptip
3y ago
We have loads more data available. For text, as LLMs start to converse constantly with users and receive direct feedback, this unlocks more and better data. Multimodal (image/video, audio) is barely being scratched. You likely could 10
930.
▲
by
theptip
3y ago
Nah, most people hadn’t thought much about AI as anything other than science fiction before ChatGPT. Non-technical people are reading the “stochastic parrot” articles, then trying a few superficial chat conversations and drawing their concl
More ›