[HN Gopher] Generative AI support on Vertex AI is now generally ...
___________________________________________________________________
 
Generative AI support on Vertex AI is now generally available
 
Author : blitz
Score  : 51 points
Date   : 2023-06-10 20:31 UTC (2 hours ago)
 
web link (cloud.google.com)
w3m dump (cloud.google.com)
 
| kumarm wrote:
| We are waiting to launch a new iOS app that has text generation
| using vertex AI for GA. So we will go live next live.
| 
| We started with GPT API but switched to Vertex AI due to speed.
| We will still use GPT API as backup still though.
 
| abusaidm wrote:
| Interesting statement and would be keen to see if businesses
| would trust Google to try out these capabilities, or other
| smaller recent services as the preferred choice given their
| flexibility of integration with existing cloud choices.
| 
| It seems we may find companies on all major cloud providers in
| the near future to guarantee access to unique proprietary
| services that cloud providers are starting to differentiate
| themselves with from their competitors
 
  | anon84873628 wrote:
  | Sure, IaaS is commodified so the next opportunity for
  | differentiation & value add is in services.
  | 
  | For GCP specifically, the Anthis/Omni stuff seems like a way to
  | sell those services even if the infrastructure isn't actually
  | in GCP.
 
| reaperman wrote:
| Extremely curious that PaLM-E, PaLI, and GPT-4 were trained to be
| multimodal (accept non-text inputs, such as images) but the
| released API's are text-only. In GCP's case, here, they've
| released PaLM-2 which is not multimodal like PaLM-E and PaLI.
| This prevents using it for visual reasoning[0].
| 
| I'm just wondering why multiple parties seem reluctant to allow
| the public to use this.
| 
| 0: https://visualqa.org
 
  | version_five wrote:
  | Presumably they're harder to censor or enforce ideological
  | constraints on. I can't see any other reason other than them
  | being worried about bad press because someone made the model do
  | something that they want to play up as bad.
 
  | lucubratory wrote:
  | The image compression/decompression from their special token
  | system wouldn't be free, it would be just as expensive as any
  | other per-pixel transformation on an image file, and it would
  | be entirely custom software doing it that they would have to
  | run on their servers. Image upload and download is a very
  | significant increase in net traffic compared to just text and
  | could make the whole venture cost a lot more. And finally, an
  | image even when downsized is going to be composed of a _lot_ of
  | tokens, so that 's going to be a lot of computational cost just
  | to run inference on it. If they haven't implemented
  | statefulness (which many haven't right now despite the
  | simplicity of the technique, field is still very new), that
  | computational cost must be repeated with every fresh API call.
  | 
  | Basically, multi-modal functionality should be an OOM increase
  | in compute, traffic, and storage requirements for anyone
  | providing it compared to a text-only model (or an only-text-
  | allowed model).
 
  | arthurcolle wrote:
  | Way too overpowered. Imagine if I can just upload images of
  | PDFs and get them to change them on the fly. So much fraud
  | instantly. To be fair, as a *prompt engineer* as a *well
  | funded* *AI startup*, it's super fun to crack apart the RLHF
  | "safety"/"alignment" modules on these models, but it's sooo
  | trivially easy that I get 100% why they aren't just opening up
  | what I call... #TheGoodStuff
 
    | samstave wrote:
    | Plus, there is a frenzy on how to maximally exploit these as
    | fast as possible from all angles, and all parties.
    | 
    | Anyone who acts all casual, as if there is not a
    | constellation of vultures circling AI right now should
    | consider themselves 'off-grid'
 
| hoschicz wrote:
| So this means now they're no longer free I suppose :(
 
| stainablesteel wrote:
| i didn't try it as this requires you to give payment information
| for a free trial and i got sidetracked
| 
| what i did learn, is that somehow, google has all of my credit
| cards despite me never sharing it on the account i was using.
 
  | Oras wrote:
  | How did you learn that Google has all your credit cards?
 
| franze wrote:
| It has been 1h.
| 
| Is it still available or has Google graveyarded it already?
 
  | samstave wrote:
  | AI moves faster than anyone could have expected.
 
| brigadier132 wrote:
| Anyone experiment with the embeddings api? How does gecko compare
| to embeddings-ada?
 
  | zetalabs wrote:
  | Where can you find gecko? Has it finally been published?
 
| williamstein wrote:
| I really really wonder how the price of vertex.so compares - in
| practice - to the openai api for use by a startup with
| unpredictable and non-sustained usage??? The multitenancy
| assumptions that are part of the openai api cost structure might
| make it much cheaper. Has anybody modeled this? I realize the
| LLM's aren't equivalent today, but longterm they could be.
 
___________________________________________________________________
(page generated 2023-06-10 23:00 UTC)