micoti

I think the reviews are flawed for most of journals. But I can't really blame them, there is huge influx of new papers comming everyday. Unfortunately reviewers are more biased towards groundbreaking stuff so it seems they can't make connections with existing and prior work.

They are also biased to see from what institution you come from..

Yes, while reading the blog I immediately thought about https://arxiv.org/html/2406.11717v3 paper. Original blog does actually overstates it findings as nothing like that exists.

Some of original blog post really seems to follow Arditi et al. (2024) (https://arxiv.org/abs/2406.11717). Projecting out a direction can only remove behaviour and not add capability so the jump mostly comes from unblocking the looping.

So I do agree it is misleading and trying to present work as groundbreaking.

Thank you very much

I am interested in this. Could you provide more info?

I know it is not related but I don't like any current available solution and also hate excell/sheets but used LLMs to handle this part for me so I can better organize my finances

Unfortunately part of some cyber sec jobs is literally doing sales. I remeber in the begging of my career I was offered cyber sec job and the guy hiring me said you will be doing sales and will be reaching out to people on top of your job. I just declined.

MacOS contstant trolling with tab switching. I regret each day of life I need to use this shitty OS because of corporate. In linux these things will never happen

Word in the Bible is not meaning the actual word (like sematic array of characters) but more like greek word logos

You are absolutely right!

i agree, these tools are usefull. i only oppose agresive marketing that llm is solution for everything. it is just a tool which has its use case, but to me it seems that it is not optimal for use cases that it is advertised.

i work on agentic systems and they can be good if agent has a bit-sized chuck of work it needs to do. problme with the coding agents is that for every more complex thing you will need to write a big prompt which is sometimes counter productive and it seems to me that user in cursor thread is pointing in that direction.

I have the exact same question, what is hype all about when models can't do simple things. You prompt the model with generate one unit test for function and it somehow always generate more then one. (Just to start with most simple instruction)

I just feel that models are currently not up to speed with experienced engineers where it takes less time to develop something then to instruct model to do it. It is only usefull for boring work.

This is not to say that these tools didn't created oportunities to create new stuff, it is just that the hype overestimates the usefullnes of the tools so they can sell them better just like all other things.