Nothing new to see here. Human also hallucinates, as you can tell from the model name.
HN user
dabei
Vibe coding a vibe coder.
Did they also delist themselves from the App Store or is that by Apple and Google?
Agree, think about it as scribbles we make on the scratch pad when thinking through a challenging math question. It mostly serve as input to our ongoing thinking process, rather than persistence of thoughts.
Yes. Similar to how "thinking step by step" helps LLMs achieve higher quality solution.
Maybe it's similar to tweaking the LLM prompts. A different set of words trigger a different pathway in our thinking.
Great read, thanks for sharing. Interesting to think about the role AI would play in theory building.
Took me a couple of minutes to realize this is not an AI launch.
Looks cool, joined the waitlist.
I read it as an experiment and hypothesis. The author is on a quest to prove their hypotheses. I appreciate the insight into the work in progress and the potential outcomes. It's more beneficial for me than waiting several months to see it fully proven.
Is there any effort to connect Scratch with LLMs?
The decision to open source is surprising at first glance. But it makes a lot of sense. With the greatly enhanced coding productivity, the ROI for tweaking open source tools is quite high. This puts closed source tools at a significant disadvantage in terms of customization. They are a little late to the AI coding assistant market, this move enables them to capture market share rapidly and slow down the competitors in building up data moat.
Very cool! There are probably a lot of other games you could integrate GPT with similar effect.
This is awesome and has huge potential to improve developer productivity! The only thing that it gives me pause is that it requires a very broad Github authorization. Do you really need to "be able to read and write all public and private repository data"?
How much does project Texas help mitigate the risk of getting banned?
Faster response time is also a big deal.
Don't ever let them find out you do it deliberately. This is quite disrespectful use of their time.
Yes, the LinkedIn recruiter account.
I haven’t used the SO recruiting tool. But I’d imagine it would let me skim the questions a candidate have been asking / answering. So I get a sense of their technical / communication skill, and passion about their domain.
As an engineering manager I found LinkedIn to be very inefficient in finding quality talents at scale. There isn’t enough signal to gauge which candidate is worth bringing in for interviews. There seem to be a large opportunity for new business here.
Great job! What makes PioSolver much more popular than the other solvers?
Maybe there is some truth here but they haven’t done a good job if that is the goal. The simplified Chinese is not that different from traditional Chinese. We can still read traditional text easily. The biggest difference is the Simplified Chinese is much easier to write (by hand).
You are right logically. However given the amount of confusion in this thread surely you can see a lot of people _feel_ offended. It’s worth thinking about how to avoid such ambiguity in the future.
I don't think they are talking about build logs. They are talking about the logs of a version of their app. One indication is they gave an example of identifying login errors.
How does this compare to Shaka packager?
Sure. I'm just interested to know what else in EXIF could be identifying.
Nice. The most identifying EXIF data is probably the GPS location. What else have to be scrubbed?
I think you are suffering from perfectionism. You are in a healthy process of continuous improvement, keep discovering things that you could have done better. But there could be thousands of engineers out there that could have benefited from being more like the version of you yesterday.
The ones who wouldn't benefit from it would just move on to read material more suitable for their level. You don't need to worry about protecting them from making the same mistakes as you do.
What are some interesting applications of edge anomaly detection?
Is this an ad?