11–20 of 41 posts
My father who doesn’t speak English well, was experiencing a heart attack at home at night by himself and he asked it symptoms and it told him to drive himself to an ER, so he listened to it. I’m thankful that he’s here today.
Because it's the wrong question!
It's not that LLMs solve entire classes of life/work problems. Instead, they take some life/work task (coding, ideas generation, learning about new topics, personal reflection) and make them x% easier, y% faster, z% better.
The industry was flooded with "talent" 2018-2022. LLM dependence has lowered the bar for professional excellence while thinning the ranks of talented newcomers by discouraging them. I think I'll be able to work through retirement age without having to settle for eastern European rates. In 2021 that seemed less likely.
If (the royal) your claim to fame is “I codez real gud”, you would be screwed post 2022 with or without LLMs.
On that same note, at 51 years old, if my only means of staying competitive and employable is that I can reverse a b tree on the whiteboard, I’ve done something horribly wrong with my life.
Now, before you jump on me saying that AI is wrong, this is true. But at the same time, I no longer can be 100% sure that whatever SEO optimized website I land at provides accurate information. If I need solid facts, I usually double check AI with various other sources. For queries like "best keyboard for software engineers", I'd rather get a table with pros/cons from AI rather than landing on whatever affiliate related website is promoted on Google. LLM gives me a good starting point to either dig deeper into particular products, or query further to find more suggestions.
Same for coding. I used to Google "how to split a string in ruby" and land on flame war, or 19 years old, StackOverflow question. Now I can get an updated answer from whatever LLM you prefer with a reference to the official documentation. It works for simple queries, as well as code snippets.
Lastly, I use LLMs to plan trips or gift ideas. I'd just throw in my preferences, and let LLM build a rough plan, from which I can iterate further, or start doing my own research.
Claude and Gemini have been very useful in helping me come up to speed on a code base written in Go (a language I have used before but not for many years). Figuring out where the business logic lies, how the dependency injection is done, how the tests are written, what overall design pattern is being etc.
Of course, I could have done all this without LLMs but it would have taken several weeks/months longer. Letting the LLM handle the boilerplate and framework jargon lets me focus on the business logic and the design patterns, and helps me contribute much faster. But LLMs do often make mistakes so it's not like I blindly trust the output. They don't replace your colleagues in terms of being the ultimate source of truth. But it has speeded up the learning process, no doubt.
Also, when writing code I provide the style guide to the LLM as context and have it review the code.
The industry was flooded with "talent" 2018-2022. LLM dependence has lowered the bar for professional excellence while thinning the ranks of talented newcomers by discouraging them. I think I'll be able to work through retirement age without having to settle for eastern European rates. In 2021 that seemed less likely.
LLMs don’t help at all with engineering and knowing what problems should be solved. Knowing how to deal with XYProblems, dealing with “the business”, go to market strategies, etc. They help with coding . If (the royal) your claim to fame is “I codez real gud”, you would be screwed post 2022 with or without LLMs. On that same note, at 51 years old, if my only means of staying competitive and employable is that I can r…
A common argument I hear about AI is "I could just write it faster myself", well I know CompSci and general info about a lot of software things but it would take hours of getting up to speed on areas I'm not an expert in to be productive. I can just delegate that to AI and get mostly correct outputs, this is okay with me and faster than what I could do.
I think the cost is going to catch up with the AI companies running the models (not the companies building products that call AI APIs) and that is when the bubble will burst. They will need to keep increasing costs and at some threshold, fewer and fewer developers in an organization will have licenses because it may become unaffordable.
Earlier quoted context omitted.
LLMs don’t help at all with engineering and knowing what problems should be solved. Knowing how to deal with XYProblems, dealing with “the business”, go to market strategies, etc. They help with coding . If (the royal) your claim to fame is “I codez real gud”, you would be screwed post 2022 with or without LLMs. On that same note, at 51 years old, if my only means of staying competitive and employable is that I can r…
Luckily you're not me. Reality check, anything that keeps you employable in a well playing job past 50 is a win.
Yes I code as part of my day job depending on the way the wind is blowing. But I get hired because I can talk to CxOs, directors and people with budget decisions on zoom or hop on a plane. Even my interview at AWS was all system design and behavioral.
If I ever responded to recruiters or people I know through my network at GCP, that’s the way I would get a job there.
But I would rather get a daily anal probe with a cactus than ever work for BigTech again and I’m damn sure not going back into an office.