The scare quotes around words that don't warrant it, or are unnecessarily idiosyncratic, are something I get pretty often in response text from Gemini.
In this case the use of quotes seems to have been perfectly appropriate as it's almost certainly a word they've seen many people using when giving feedback.
I'm really surprised that didn't jump out at more people; I had to get halfway through the comments to the 27th mention of "Department of War" to find the first comment pointing out that using the name is itself a capitulation.
Defense is a much more fitting name for an organization that does a million more things than just prosecute wars. War is just the favorite part of their mission for these wannabe toughguys.
For me there seems to be a listing of configurable settings or something but it only pops up for a single frame after I right-click-drag a component - this seems like a broken mouse interaction.
I strongly suspect there's a major component of this type of experience being that people develop a way of talking to a particular LLM that's very efficient and works well for them with it, but is in many respects non-transferable to rival models. For instance, in my experience, OpenAI models are remarkably worse than Google models in basically any criterion I could imagine; however, I've spent most of my time using the Google ones and it's only during this time that the differences became apparent and, over time, much more pronounced. I would not be surprised at all to learn that people who chose to primarily use Anthropic or OpenAI models during that time had an exactly analogous experience that convinced them their model was the best.
While this is certainly very true, I find coding through an LLM to require far less effort dedicated to this cognitive switching than does writing in some programming language, primarily because I no longer have to load the mental context for converting my high level human instructions to code that a programming environment actually supports. The mental context seems more lightweight and closer to the way I think about the problem when I'm not sitting at the computer actively working on it. If an idea comes to me while I'm away from the computer I can momentarily sit down, type in whatever I just thought of, and get going almost immediately. I think it also saves a huge amount of cognitive load and stress (for me) involved with switching around between different programs and languages, an unfortunate fact of life when dealing with legacy systems.
I just kept scrolling, hoping it would learn from how long I paused over content to read it the way FB's seems to, but it seems you're right, in this case "likes" are required.
It's really a stretch for the article to suggest their gear might not have supported fractional bpm. MIDI itself has always supported it and analog sequencers before that support it even easier. Not to mention external clock sync has been a thing for decades.
For me it just depends. If the response to my prompt shows the model misunderstood something, then I go back and retry the previous prompt again. Otherwise the "wrong ideas" that it comes up with persist in the context and seem to sabotage all future results. The most of this sort of coding I've done was in Google's AI studio, and I often do have a context that spans dozens of messages, but I always rewind if something goes off-track. Basically any time I'm about to make a difficult request, I clone the entire context/app to a new one so I can roll back [cleanly] whenever necessary.
If you fix something it sticks, the AI won't keep making the same mistake, it won't change the code that already exists if you ask it not to. It actually ONLY works well when you are doing iterative changes and not used as a pure code generator, actually, AI's one-shot performance is kind of crap. A mistake happens, you point it out to the LLM and ask it to update the code and the instructions used to create the code in tandem. Or you just ask it to fix the code once. You add tests, partially generated by the AI and curated by a human, the AI runs the tests and fixes the code if they fail (or fixes the tests).
All I can really say is that doesn't match my experience. If I fix something that it implemented due to a "misunderstanding" then it usually tends to break it again a few messages later. But I would be the first to say the use of these models is extremely subjective.
I think we have very different experiences then. I find multiple prompts with narrow focuses each executed to update the same file work much better than trying to one shot the file. I think you would have a better experience if you used /clear (assuming you are using Gemini CLI), the problem isn't the change in the file, the problem is probably the conversation history instead.
You need talented people to turn bad publicity into good publicity. It doesn't come for free. You can lose a lot with a bad rep.
Those talented people that work on public relations would very much prefer working with base good publicity instead of trying to recover from blunders.
"Open your mouth and say ah" "tot" "yacht" - these all have very close to the same vowel sound to me as an American, although "tot" is more of an outlier and "taught" might be closer to how I conceptualize of the sound. I'm not sure I'd ever hear the difference in practice.
reply