msvana

I think there is one problem with defining acceptance criteria first: sometimes you don't know ahead of time what those criteria are. You need to poke around first to figure out what's possible and what matters. And sometimes the criteria are subjective, abstract, and cannot be formally specified.

Of course, this problem is more general than just improving the output of LLM coding tools

I work as an ML engineer/researcher. When I implement a change in an experiment it usually takes at least an hour to get the results. I can use this time to implement a different experiment. Doesn't matter if I do it by hand or if I let an agent do it for me, I have enough time. Code isn't the bottleneck.

I also heard an opinion that since writing code is cheap, people implement things that have no economic value without really thinking it through.

This reminds me of the idea that LLMs are simulators. Given the current state (the prompt + the previously generated text), they generate the next state (the next token) using rules derived from training data.

As simulators, LLMs can simulate many things, including agents that exhibit human-like properties. But LLMs themselves are not agents.

More on this idea here: https://www.alignmentforum.org/posts/vJFdjigzmcXMhNTsx/agi-s...

This perspective makes a lot of sense to me. Still, I wouldn't avoid anthropomorphization altogether. First, in some cases, it might be a useful mental tool to understand some aspect of LLMs. Second, there is a lot of uncertainty about how LLMs work, so I would stay epistemically humble. The second argument applies in the opposite direction as well: for example, it's equally bad to say that LLMs are 100% conscious.

On the other hand, if someone argues against anthropomorphizing LLMs, I would avoid phrasing it as: "It's just matrix multiplication." The article demonstrates why this is a bad idea pretty well.

I had this exact analogy in mind when writing point no. 3

Thanks, sounds kinda similar to my first point.

Problem no. 2 (Understanding user intent) is relevant not only to writing SQL but also to software development in general. Follow-up questions are something I had in mind for a long time. I wonder why this is not the default for LLMs.

At the beginning, the article mentions correlation with language skills AND problem-solving. Focusing only on language skills in the second half is misleading. According to the abstract of the original paper, problem solving and working memory capacity were FAR MORE important.

Also, the article doesn't mention "math skills". It talks about numeracy, which is defined in a cited paper as "the ability to understand, manipulate, and use numerical information, including probabilities". This is only a very small part of mathematics. I would even argue that mathematics involves a lot of problem solving and since problem solving is a good predictor, math skills are good predictor.

Hmm this gave me an interesting project idea: a coding assitant that talks shit about your lack of skills and low code quality.

Hopefully you'll forgive my ignorance, but this is the first time I hear about DuckDB. What space does it occupy in the DBMS landscape? What are its use cases? How does it compare to other DBMS solutions?

I have a few thoughts after reading this:

- I started to see LLMs as a kind of search engines. I cannot say they are better than traditional search engines. On one hand, they are better at personalizing the answer, on the other hand, they hallucinate a lot.

- There is a different view on how new scientific knowledge is made. It's all about connecting existing dots. Maybe LLMs can assist with this task by helping scientists discover relevant dots to connect. But as the author suggests, this is only part of the job. To find the correct ways to connect the dots, you need to ask the right questions, examine the space of counterfactuals, etc. LLMs can be useful tool, but they are not autonomous scientists (yet).

- As someone developing software on top of LLMs, I am slowly coming to a conclusion that human-in-the-loop approaches seem to work better than fully autonomous agents.

Well I guess the author of the blog post agrees, since he talks about "its demise in the 1980s". Probably a lot of things have changed since then and honestly I am kinda curious about it.

I don't think using your real name helps that much. Most people on Facebook use their real name and still many have no issues with insulting other people or even supporting violence against individuals or groups of people.

You are capable of dealing with most bad events. Very few situations are truly life-ruining.

Not a big fan of Ubuntu since they started forcing snaps on their users. I like many distros based on Ubuntu though. Like Pop!_OS. It was my daily driver for quite some time. But half a year ago I switched to Fedora mainly out of curiosity. The user experience isn't that different. Both distros use GNOME and have Flatpak.

Sort of a meta answer: learning how to learn. It feels like the velocity of change is pretty high. So being able to pick up new skills quickly is important.

I also agree with others that soft skills are incredibly important.

As for fields of expertise, I am looking into many diverse topics such as urban planning, sociology, civil engineering, aerospace engineering, electricity production. I feel like at least in my environment there are a lot of problems that I could help solve if I had knowledge in some of these areas. Besides that I am doing a PhD and recently there has been a lot of emphasis on interdisciplinary research.