@neetcode1@vasilios_s A discussion from which people actually have something to gain, especially regarding expectations, learning, and setting and achieving goals.
When you write code with AI, the rules of software engineering do not suddenly disappear. A developer can turn code into unmaintainable mess in months, AI can do it in days. Just like it always was: if you need maintainable code, pay a lot of attention to architecture and tests.
"Orion is built by just six developers (for the first two years it was just one developer) and doesn’t use ad revenue or data sales."
https://t.co/AaHWoCKbWy
Our small but mighty team 👇
We are looking for a talented and detail-oriented developer to join our marketing team's efforts in creating and maintaining our website, product landing pages, newsletter templates & more. Fully remote.
For details & to apply: https://t.co/ZEtTRqk9W7
CV's are full of shit. All those impact summaries are all too often exaggerations, and who knows how much of a project listed someone actually did? No, if you want the truth, you gotta look at real work.
A key finding in this work: adding irrelevant context to GSM8k problems causes LLMs to fail at solving them, as we demonstrated in our ICML 2023 paper, "Large Language Models Can Be Easily Distracted by Irrelevant Context" (https://t.co/WY85ibOttc). The differences in prompt construction still look interesting to me.