Right, based off this phrasing I was expecting USD to be at 5-year lows compared to JPY, CHF, EUR, RMB, BRL, etc. but that does not seem to be the case
The real title of the post seems too obscure. The one I picked ("The world doesn't have enough money to pay for this much AI") is a quote from the article and captures the main idea more directly.
* Writing style. In agent docs, using all caps might be an effective way to emphasize a particular instruction. In internal eng docs, this might come off rude or distracting.
* Conciseness vs. completeness. In agent docs, you likely need to keep the content highly curated. If you put in too much content, you’ll blast through your API quotas quickly and will probably reduce LLM output quality. In internal eng docs, we ideally aim for 100% completeness. I.e. every important design decision, API reference, workflow, etc. is documented somewhere.
* Differing knowledge needs. The information that LLMs need help with is not the same as the information that human engineers need help with. For example, Gemini 2.5 Pro has pretty good built-in awareness of Pigweed’s C++ Style Guide. I tested that assertion by invoking the Gemini API and instructing it Recite the Pigweed C++ Guide in its entirety. It did not recite in full, but it gave a detailed summary of all the points. So the Gemini 2.5 Pro API was either trained on the style guide, or it’s able to retrieve the style guide when needed. Therefore, it’s not necessary to include the full style guide as AGENTS.md context. (Credit to Keir Mierle for this idea.)
Arguments against:
* Duplication. Conceptually, agent docs are a subset of internal eng docs. The underlying goal is the same. You’re documenting workflows and knowledge that’s important to the team. But now you need to maintain that same information in two different doc sets.
> Writing style. In agent docs, using all caps might be an effective way to emphasize a particular instruction. In internal eng docs, this might come off rude or distracting.
To pile on to this, an agent needs to see "ABSOLUTELY NEVER do suchandsuch" to not do suchandsuch, but still has a pretty fair chance of doing it by accident. A talented human seeing "ABSOLUTELY NEVER do suchandsuch" will interpret this to mean there are consequences to doing suchandsuch, like being fired or causing production downtime. So the same message will be received differently by the different types of readers.
This. When doing Stable Diffusion, I have noticed this as well. Adding negatives can sometimes lead to the opposite results.
From what I can tell, if you say "no computers" for example (ie adding computer as negative), you are setting the scene for something like "where there should be computer, there is not".
I can't better describe this phenomenom, only that it can completely change the output in unexpected unwanted ways.
I can, I don't have a specific example I've used to give you in this moment. And trying to share an exact example would read like a double negative.
The general rule of thumb is only put what you want in context. If you put instructions of what not to do in context, those tokens can be misunderstood and create unintended/unwanted steering of the model.
A fair example would be testing for positive sentiment. Consider weight of tokens appended to context, phrase instructions or questions to be neutral or positive.
e.g. Some phrases and their impact:
- "Is the tone of the user message positive?" will be biased for a false positive.
- "Analyze the tone of the user message?" will be more neutral and less biased.
- "Is the tone of the message negative?" will be biased for false positives when evaluating for negative tone.
I have tried using a GitHub repo to host photos that I displayed on a different website. IIRC it didn't work great. I would try to access the photos over the raw GitHub URLs and I'm pretty sure they would often 404. Was I holding it wrong? Are there any great guides on this topic? I also remember that "uploading" photos over Git was a pain. Basically could only upload one at a time.
I meant using GitHub to host your whole site through gh-pages, not hotlinking to assets from some other unrelated server. You can even use your own domain.
Quite badly. Can't tell you how many times an LLM has suggested WORKSPACE solutions to my Bazel problems, even when I explicitly tell them that I'm using Bzlmod.
Graduated with a B.A. in History from Cal in 2011. Struggled for 6+ months to land a job. Took any miscellaneous writing job on oDesk (now Upwork) that I could get. Saw a lot of contracts for something called "technical writing" and they often paid much better than other writing work. Figured out that it was basically "instruction manuals for computer professionals". Realized it was a good fit. Took a bunch of basic C.S. courses at community college. Family friend got my foot in the door at an IoT startup, but it was really those C.S. classes that persuaded them to give me a try. And the fact that I would work for peanuts. Turned out that I was competent in the work and I really enjoyed programming. A few years later, Google recruited me, and I've been at Google ever since.
I technically have an A.S. in C.S. from that community college, but I think my circuitous path fits the "non-traditional STEM background" criteria.
Computers have been great to me! Very thankful that I stumbled into this path.
