> I get the feeling GCP is not good for individuals like I.
Google isn't good for individuals at all. Unless you've got a few million followers or get lucky on HN, support is literally non-existent. Anyone that builds a business on Google is nuts.
I'd like to state the AWS, in contrast, has been great to me as an individual. The two times that I needed to speak to a human, I had one on the phone resolving my issue. And both issues were due to me making mistakes - on my small personal account.
Yes, it’s extremely complicated. I gave up on fire base for one project because I could not figure out how to get the right permissions set up and my support request resulted in someone copying and pasting a snippet from the instructions that I obviously had not understood in the first place.
It’s also extremely cumbersome to sign up for Google AI. The other day I tried to get deep seek working via Google’s hosted offering and gave up after about an hour. The form just would not complete without error and there was not a useful message to work with.
It would seem that in today’s modern world of AI assistance, Google could set up one that would help users do the simplest things. Why not just let the agent direct the user to the correct forms and let the user press submit?
Oh man, I've been playing with GCP's vertex AI endpoints, and this is so representative of my experience. It's actually bananas how difficult it is, even compared to other GCP endpoints
At 0:52 in their demo video, there is a grammatical inconsistency in the agent's text output. The annotations in the video are therefore suspected to be created by humans after the fact. Is Google up to their old marketing/hyping tricks again?
> SIMA 2 Reasoning:
> The user wants me to go to the ‘tomato house’. Based on the description ‘ripe tomato’, I identify the red house down the street.
The scene just before you describe has the user write "ripe tomato" in the description - you can see it in the video. The summary elides it, but the "ripe tomato" instruction is also clearly part of the context.
Same. I think there’s an untapped market (feature really) here, which if isn’t solved by GPT-next will start to reveal itself as a problem more and more.
LLMs are really bad at being comprehensive, in general, and from one inference to the next their comprehensive-ness varies wildly. Because LLMs are surprising the hell out of everyone with their abilities, less attention is paid to this; they can do a thing well, and for now that’s good enough. As we scale usage, I expect this gap will become more obvious and problematic (unless solved in the model, like everything else).
A solution I’ve been toying with is something like a reasoning step, which could probably be done with mostly classical NLP, that identifies constraints up front and guides the inference to meet them. Like a structured output but at a session level.
I am currently doing what you suggest though, I have the agent create a script which invokes … itself … until the constraints are met, but that obviously requires that I am engaged there; I think it could be done autonomously, with at least much better consistency (at the end of the day even that guiding hand is inference based and therefore subject to the same challenges).
> if LLMs "knew" when they're out of their depth, they could be much more useful.
I used to think this, but no longer sure.
Large-scale tasks just grind to a halt with more modern LLMs because of this perception of impassable complexity.
And it's not that they need extensive planning, the LLM knows what needs to be done (it'll even tell you!), it's just more work than will fit within a "session" (arbitrary) and so it would rather refuse than get started.
So you're now looking at TODOs, and hierarchical plans, and all this unnecessary pre-work even when the task scales horizontally very well (if it just jumped into it).
reply