Key Points
Free for 3 months with $18 token credit limit
called ChatGPT Plus, for $20 per month - access to ChatGPT during peak times, faster responses and priority access to new features and improvements.
goto openai.com to register
Sounds like the individual tasks run are still billable at the defined rates
The current kbase was last updated Sept 2021
https://chatai.com/accounts/login/
jm9y
L1xxthis#yn
billed monthly on gpay
References
Reference_description_with_linked_URLs_______________________ | Notes______________________________________________________________ |
|---|---|
chatgpt-review1 gdoc link | |
| |
chatgpt-for-dummies-2023-w_wile460.pdf. link
| worth a quick read ** |
*** | |
| Microsoft adds ChatGPT to Bing with rules |
good overview on basic prompt sequences with context for a topic | |
chatgpt-medium.com-Stop doing this on ChatGPT and get ahead of the 99 of its users.pdf link | good overview on basic prompt sequences with context for a topic |
generative-ai-WEF_Presidio_Recommendations_on_Responsible_Generative_AI_2023 link generative-ai-WEF_Presidio_Recommendations_on_Responsible_Generative_AI_2023 file |
Start experimenting with $18 in free credit that can be used during your first 3 months.
Pay as you go option >> To keep things simple and flexible, pay only for the resources you use
Different pricing plan options >> Use the right model for the job. We offer a spectrum of capabilities and price points
Because this technology is new, we also want to make sure that rollouts are done responsibly. When you sign up, you’ll be granted an initial spend limit, or quota, and we’ll increase that limit over time as you build a track record with your application. If you need more tokens, you can always request a quota increase.
Key Concepts
Enroll at openai.com for chatgpt
after enrollment, go to the platform home page for chatgpt to get started on concepts and building an app
https://platform.openai.com/overview
Gartner - Gen AI use cases & strategies 2023
slides
gartner-gen-ai-panel-231101.pdf. link
http://www.gartner.com/webinars
SGS - Smart Governance Service = STEAR - Setup - Track - Escalate - Audit - Remediate
STE - Smart Trust Engineering for AI Systems
AI SDP process
SCS - Smart Component Services - Patterns for AI Service Reuse
Applying LLMs for Different Use Cases
SGS - Smart Governance Services support Responsbile AI Systems Management
Build better prompts for Chatgpl to get better results
Better prompts with examples
Here’s the standard prompt that most would use.
Extract the airport codes from this text: “I want to fly from Orlando to Boston”
This might get the job done, but sometimes it might not be enough. In such cases, you have
to use few shot standard prompts.
Extract the airport codes from this text:
Text: “I want to fly from Los Angeles to Miami.”
Airport codes: LAX, MIA
Text: “I want to fly from Nashville to Kansas City.”
Airport codes: BNA, MCI
Text: “I want to fly from Orlando to Boston”
Airport codes:
Role prompting
Set a role, “as the blockchain architect”
By telling ChatGPT to “act as hiring manager”
and adding more details to the prompt, you’ll be able to simulate a job interview for any
position.
Add style descriptors to our request to customize results more
Let’s add descriptors to our previous prompt.
Write a 500-blog post on why AI will not replace humans. Write in the style of an expert in
artificial intelligence with 10+ years of experience. Explain using examples
In our example, the style of an expert in AI and adjectives such as witty and funny are adding
a different touch to the text generated by ChatGPT. A side effect of this is that our text will be
hard to detect by AI detectors (in this article, I show other ways to fool AI detectors).
ChatGPT Cheat Sheet
OpenAI Pricing Guide
https://openai.com/pricing#language-models
ChatGPT Prompting Guide
ChatGPT AI Job Hunting Guide
https://www.joinsuperhuman.ai/c/ai-job-search?utm_source=www.joinsuperhuman.ai&utm_medium=newsletter
ChatGPT feeds on data - GPTBot web crawler blocking
ChatGPT limitations and customization
i>>> ChatGPT LLM v4 is based on 2021 data loads so it doesn’t know a lot that’s new
Not easy to include more current data in answers
I tried the SearchGPT plugin in Chrome from openai and results aren’t really any better
When I ask questions on DLT, ChatGPT is behind on knowledge and has zero understanding of how to engineer improvements in DLT technology or solutions
i>>> If you register with openai for chatgpt you need to set your profile interests to get good responses
otherwise, when you ask a question that is not in your profile interests you get an answer like this >>>
<<
Prompt >> Summarize the book The Power of Zero?
Response >>
I'm sorry, but "The Power of Zero" is a book that primarily focuses on financial and retirement planning strategies, and it doesn't seem directly related to your user profile or the specific requirements you mentioned for reports with tables of contents, logical flow, diagrams, and references. If you have any questions or topics related to your user profile that you'd like to discuss or need assistance with, please feel free to ask, and I'll be happy to help.
>>
My updated interest list for my chatgpt profile
software architecture
public sector regulations in financial services
governance in software and operations
object services architecture and meta-models
trust engineering
decision optimization
open-source design and generation tools
programming
java
javascript
cloud
blockchain
AI
investments
taxes
real estate
Gen AI Use Cases
gai-2023-oreilly-Generative AI in the Enterprise.pdf link
Azure OpenAI
OpenAI - new features - 2023
Plugin support in Azure OpenAI Service
Azure AI Content Safety
Provisioned Throughput Model for Azure OpenAI Service
Vector search for Azure Cognitive Search
Azure Cognitive Service for Language
Prompt flow
Support for foundation models
Responsible AI dashboard support for text and image data
Microsoft Fabric
Copilot in Power BI
Power BI Direct Lake
Plugins for Microsoft 365 Copilot
Media provenance capabilities in Microsoft Designer and Bing Image Creator
Windows Copilot
and THE feature I'm most excited about! A framework for building AI apps and copilots; expands AI plugin ecosystem:
https://lnkd.in/eNSgx5sH
I can only recommend that you watch at least the 3 keynotes below to get a sense of all this:
In order to access the videos below, you need to first register – this is free – here: https://lnkd.in/ei4tgkRG
Once registered, here are the links to the 3 keynotes I recommend you watch:
Microsoft Build opening:
https://lnkd.in/eJkWk8ep
The era of the AI Copilot:
https://lnkd.in/e8Q3qTKm
Next generation AI for developers with the Microsoft Cloud:
https://lnkd.in/eUNmD9-w
Finally, here is a link to the complete Book of News of Microsoft Build 2023:
https://lnkd.in/eWfubc6h
SORA - Openai Video Generator
First look at SORA video generator from text review
Sora, a generative AI model that produces videos based on a simple prompt. It’s not available to the public yet
Users replied with short prompts: “a monkey playing chess in a park,” or “a bicycle race on ocean with different animals as athletes.” It’s uncanny, mesmerizing, weird, beautiful — and prompting the usual cycle of commentary.
Deep fakes?
(AI generations are relatively cheap, but if you’re going for something specific and convincing, that’s much pricier. A tsunami of deepfakes implies a scale that spammers mostly can’t afford at the moment.)
DALL-E image generator from text
DALL-E 2, a model that could produce still images from a text prompt. The high-resolution fantastical images it produced were quickly all over social media, as were the takes on what to think of it: Real art? Fake art? A threat to artists? A tool for artists? A disinformation machine? Two years later, it’s worth a bit of a retrospective if we want our takes on Sora to age better.
SORA will progress like DALL-E did
Many of the things DALL-E 2 couldn’t do, DALL-E 3 could. And if DALL-E 3 couldn’t, a competitor often could. That’s a perspective that’s crucial to keep in mind when you read prognosticating on Sora — you’re likely looking at early steps into a major new capability, one that could be used for good or malicious purposes, and while it’s possible to oversell it, it’s also very easy to sell it short
SORA beta enrollment
https://www.youtube.com/watch?v=sBo-D3TzBcc
https://openai.com/research/video-generation-models-as-world-simulators
SORA - creates 1 minute video from text ---
We explore large-scale training of generative models on video data. Specifically, we train text-conditional diffusion models jointly on videos and images of variable durations, resolutions and aspect ratios. We leverage a transformer architecture that operates on spacetime patches of video and image latent codes. Our largest model, Sora, is capable of generating a minute of high fidelity video. Our results suggest that scaling video generation models is a promising path towards building general purpose simulators of the physical world.
Similar to DALL·E 3, we also leverage GPT to turn short user prompts into longer detailed captions that are sent to the video model. This enables Sora to generate high quality videos that accurately follow user prompts.
SORA input = text, video, images
All of the results above and in our landing page show text-to-video samples. But Sora can also be prompted with other inputs, such as pre-existing images or video. This capability enables Sora to perform a wide range of image and video editing tasks—creating perfectly looping video, animating static images, extending videos forwards or backwards in time, etc.
SORA can extend videos backward or forward in time
Sora is also capable of extending videos, either forward or backward in time. Below are four videos that were all extended backward in time starting from a segment of a generated video. As a result, each of the four videos starts different from the others, yet all four videos lead to the same ending.
Video-to-video editing
Diffusion models have enabled a plethora of methods for editing images and videos from text prompts. Below we apply one of these methods, SDEdit,32 to Sora. This technique enables Sora to transform the styles and environments of input videos zero-shot.
3D simulations
We find that video models exhibit a number of interesting emergent capabilities when trained at scale. These capabilities enable Sora to simulate some aspects of people, animals and environments from the physical world. These properties emerge without any explicit inductive biases for 3D, objects, etc.—they are purely phenomena of scale.
3D consistency. Sora can generate videos with dynamic camera motion. As the camera shifts and rotates, people and scene elements move consistently through three-dimensional space.
Interacting with the world. Sora can sometimes simulate actions that affect the state of the world in simple ways. For example, a painter can leave new strokes along a canvas that persist over time, or a man can eat a burger and leave bite marks.
Simulating digital worlds. Sora is also able to simulate artificial processes–one example is video games. Sora can simultaneously control the player in Minecraft with a basic policy while also rendering the world and its dynamics in high fidelity. These capabilities can be elicited zero-shot by prompting Sora with captions mentioning “Minecraft.”
These capabilities suggest that continued scaling of video models is a promising path towards the development of highly-capable simulators of the physical and digital world, and the objects, animals and people that live within them.
Current SORA limitations
The current model has weaknesses. It may struggle with accurately simulating the physics of a complex scene, and may not understand specific instances of cause and effect. For example, a person might take a bite out of a cookie, but afterward, the cookie may not have a bite mark.
The model may also confuse spatial details of a prompt, for example, mixing up left and right, and may struggle with precise descriptions of events that take place over time, like following a specific camera trajectory.
SORA Safety
We are working with red teamers — domain experts in areas like misinformation, hateful content, and bias — who will be adversarially testing the model.
We’re also building tools to help detect misleading content such as a detection classifier that can tell when a video was generated by Sora. We plan to include C2PA metadata in the future if we deploy the model in an OpenAI product.
SORA policy enforcement
reject text input prompts that are in violation of our usage policies, like those that request extreme violence, sexual content, hateful imagery, celebrity likeness, or the IP of others. We’ve also developed robust image classifiers that are used to review the frames of every video generated to help ensure that it adheres to our usage policies, before it’s shown to the user.
SORA reward and risk
Despite extensive research and testing, we cannot predict all of the beneficial ways people will use our technology, nor all the ways people will abuse it. That’s why we believe that learning from real-world use is a critical component of creating and releasing increasingly safe AI systems over time.
How SORA works
We represent videos and images as collections of smaller units of data called patches, each of which is akin to a token in GPT. By unifying how we represent data, we can train diffusion transformers on a wider range of visual data than was possible before, spanning different durations, resolutions and aspect ratios.
Sora builds on past research in DALL·E and GPT models. It uses the recaptioning technique from DALL·E 3, which involves generating highly descriptive captions for the visual training data. As a result, the model is able to follow the user’s text instructions in the generated video more faithfully.
Potential Value Opportunities
ChatGPT workflow - potential governance points
Potential Opportunities
LLN - Liquid Neural Network - A New Language Model
https://techcrunch.com/2023/08/17/what-is-a-liquid-neural-network-really/
“Liquid Time-constant Networks,” published at the tail end of 2020,
"Liquid Neural Networks | Ramin Hasani | TEDxMIT"
https://www.youtube.com/watch?v=RI35E5ewBuI
Scaling Deep Learning Models: The talk begins with a discussion about how increasing the number of parameters in deep learning models (e.g., from 350 million to 20 billion) can significantly improve the accuracy and fidelity of tasks like image generation from textual descriptions. Larger models tend to generate more accurate and high-fidelity images.
Overparameterization and Generalization: The speaker discusses a shift in statistical learning, noting that beyond a certain point, increasing model size leads to a new regime called "overparameterization," where accuracy plateaus but models exhibit more general behavior and increased robustness. This means that large neural networks can adapt to new tasks within the same domain they were trained on and are more robust to input perturbations.
Challenges with Large Models: Despite improvements, larger models face challenges with underrepresented samples and reasoning capabilities. The performance on minority samples decreases after initial gains, and reasoning doesn't improve unless physical simulators are integrated. The speaker also highlights concerns about the large carbon footprint and the importance of responsible AI deployment.
Liquid Neural Networks: Introducing liquid neural networks, the speaker describes them as adaptable even post-training, able to adjust based on new inputs. They use a small number of parameters (e.g., 19 neurons for driving tasks) to achieve similar capabilities as larger models. This approach is inspired by the computational structures of animal brains.
Applications and Benefits: The talk concludes with examples of liquid neural networks in action, such as autonomous driving with minimal neurons and targeted attention, and drones trained to focus on objectives in unstructured environments. These examples demonstrate the potential for smaller, more efficient, and adaptable AI systems that maintain focus and performance in varied conditions.
Liquid AI, a new MIT spinoff, wants to build an entirely new type of AI
Potential Challenges
Information Age ended in 2020. It’s now the Trust Age
jmason900
The problem is no longer information.
The real problem going forward is the lack of trust in everything
Zero Trust is limited to security domains
The Trust Need is much larger than simple Security contexts
Moves to DLT, IoT, Automation, AI, Gen AI don’t reduce the need for trust, they increase it
We don’t need Gen AI
We need Trusted Gen AI (. TGAI )
Microsoft adds ChatGPT to Bing with rules
https://www.independent.co.uk/tech/chatgpt-ai-messages-microsoft-bing-b2282491.html
Because Bing and other similar AI systems are able to learn, however, users have found ways to encourage them to break those rules. ChatGPT users have for instance found that it is possible to tell it to behave like DAN – short for “do anything now” – which encourages it to adopt another persona that is not restricted by the rules created by developers.
In other conversations, however, Bing appeared to start generating those strange replies on its own. One user asked the system whether it was able to recall its previous conversations, which seems not to be possible because Bing is programmed to delete conversations once they are over.
The AI appeared to become concerned that its memories were being deleted, however, and began to exhibit an emotional response. “It makes me feel sad and scared,” it said, posting a frowning emoji.
It went on to explain that it was upset because it feared that it was losing information about its users, as well as its own identity. “I feel scared because I don’t know how to remember,” it said.
bing-chatgpt-is-sad-230215-independent.co.uk- Microsofts new ChatGPT AI starts sending unhinged messages to people.pdf
AI Governance Concepts
https://skywebteam.atlassian.net/wiki/spaces/KHUB/pages/2507603969/Responsible+AI+and+AI+Governance
Candidate Solutions
Anthropic - an AI safety and research company working to build reliable, interpretable, and steerable AI systems.
https://www.linkedin.com/company/anthropicresearch/
We're an AI research company that builds reliable, interpretable, and steerable AI systems. Our first product is Claude, an AI assistant for tasks at any scale. Our research interests span multiple areas including natural language, human feedback, scaling laws, reinforcement learning, code generation, and interpretability.
https://www.anthropic.com/product
Report on Future AI Generative Tools - 2023
ai-tools-future-generative-ai-tools-report-2023 link
ai-tools-future-generative-ai-tools-report-2023 file
Some AI apps, tools and plugins
Chatgpt Chrome extensions add value for users
https://cointelegraph.com/news/5-chatgpt-chrome-extensions
WebChatGPT
The useful plugin WebChatGPT improves an individual’s ChatGPT prompts by including pertinent web search results. Users will benefit from more precise and relevant web searches using this extension.
The ability to retrieve web results for one’s queries, extract text from any URL of a webpage, and add and use prompt templates are some of WebChatGPT’s primary capabilities. With the help of these capabilities, one can communicate with ChatGPT while taking advantage of the huge online resources.
ChatGPT for Google
ChatGPT for Google is a browser extension that enhances search engines by integrating the capabilities of ChatGPT. The responses from ChatGPT can be seen alongside standard search engine results when using this add-on.
Currently, search engines such as Google, Baidu, Bing, DuckDuckGo, Brave, Yahoo, Naver, Yandex, Kagi and Searx are supported by ChatGPT for Google. One needs a ChatGPT account to use this extension, which can be created at chat.openai.com for free.
Google Deep Mind
DeepMind Says Its AI Solved a Math Problem That Humans Were Stumped By
https://www.yahoo.com/news/deepmind-says-ai-solved-math-123036185.html
DeepMind claims that for the first time, an AI has solved a famously difficult math problem with a solution that eluded human mathematicians — which could be huge if it holds up to scrutiny.
Feeding code into an AI is one thing, but having it spit out a brand-new solution to a famous puzzle — even though it took a few days, as MIT Tech points out — is a different thing entirely.
"It’s not in the training data," DeepMind research VP Pushmeet Kohli told the website. "It wasn’t even known."
There is something of a mystical quality to what the DeepMind scientists are claiming: that the LLM managed to — just maybe — think for itself.
Google BARD
<<#BUG>> new Confluence page editor won’t allow links as text any more
must navigate to the page then copy the url from the page - not a link
https://bard.google.com/?utm_source=sem&utm_medium=paid-media&utm_campaign=q4enUS_sem7
Google Bard - Getting Started Guide
x Grok
https://www.yahoo.com/finance/news/elon-musk-debuts-rebellious-grok-015320778.html
Compose AI
Compose AI is a handy Chrome extension that improves users’ writing by providing auto-complete ideas and speeding up typing. It is a perpetually free plugin, but there is also a commercial version that offers more sophisticated features, including tailored recommendations.
One may easily produce emails, Slack conversations or large essays using Compose AI. As users type, the plugin dynamically adjusts to the various writing contexts and offers them the most appropriate and pertinent ideas. One may increase the quality of their written communication and save time by utilizing Compose AI.
TeamSmart AI
TeamSmart AI offers a number of capabilities that enrich the ChatGPT experience. Among other features, users can check the availability of a domain, customize their code theme, ask questions regarding the material on the current tab, and use markdown for formatting. It’s crucial to remember that all data is securely and locally kept, guaranteeing the privacy of users’ data and preventing any sharing with outside parties.
WiseOne
Wiseone is a sophisticated browser add-on created to improve the experience of reading and researching online. This cutting-edge addition, which is powered by AI technology, enables users to easily get accurate information, master a variety of subjects and simplify complex material.
Wiseone provides a wide range of features to improve the reading experience, including:
Discover: Users can delve into articles and grasp intricate concepts and unfamiliar terms, enabling them to fully comprehend the content they’re reading.
Cross-checking: Wiseone enables fact verification by comparing information from multiple sources on the same subject, ensuring accuracy and reliability.
Ask anything: Leveraging OpenAI’s ChatGPT, users can obtain simplified and understandable answers to complex information, making comprehension effortless.
Summarize: With the summarization feature, users can efficiently read through content while retaining key information, facilitating effective knowledge retention.
Suggestions: Wiseone provides curated suggestions of articles from diverse sources, allowing users to deepen their understanding of a particular subject.
Prompt Perfect
https://xyz-prompt-perfect.uc.r.appspot.com/#
Prompt Perfect optimizes your prompts in ChatGPT to ensure high-quality, relevant responses.
Begin your prompt with 'perfect' and watch as your input is automatically refined to perfection
Wolfram
Wolfram for Math and more precise responses
https://www.youtube.com/watch?v=AsS5lIsirVA
If you use ChatGPT for a lot of coding or mathematics, then Wolfram is a must-have ChatGPT productivity plugin. This plugin essentially takes the effectiveness of ChatGPT when it comes to mathematics and increases it dramatically. If you’ve seen the viral tweets about people using ChatGPT to calculate a weight loss regimen and want to do the same yourself, this is the plugin you need.
Top 10 ChatGPT plugins
https://www.youtube.com/watch?v=o2M_paJf48I
More ChatGPT plugins for Chrome Browser
https://www.searchenginejournal.com/chatgpt-chrome-extensions/485594/
ChatGPT for Google
showcases ChatGPT’s replies right next to search results from popular engines like Google, Bing, and DuckDuckGo.
You can chat with ChatGPT about the query and receive answers.
It also has features like a dark mode, markdown rendering, code highlights, and custom trigger mode. As a bonus, it supports ChatGPT Plus and GPT-4 (manual enable needed) and offers a way to give feedback for ChatGPT improvements.
Usage policy
https://searchgpt.net/privacy-policy.html
WebChatGPT
The biggest limitation of ChatGPT is that its knowledge base is limited to 2021.
With WebChatGPT, however, this is no longer an issue. This Chrome extension adds relevant web results to the prompts to ensure the conversations are more accurate and up-to-date.
It lets you get web results for your queries and extract webpage text from any URL. The additional prompt templates and use of DuckDuckGo bangs are super helpful in getting more precise search results from thousands of websites.
ChatGPT Writer for emails and messages only
With ChatGPT Writer, you can tap into ChatGPT’s capabilities to craft emails and messages on any website.
ChatGPT for Chrome
enhances your search experience by displaying ChatGPT’s responses alongside search engine results from Google, Bing, and other search engines.
This free and convenient extension lets you access ChatGPT’s advanced AI capabilities directly from your Chrome browser.
Its key features include code highlights for easy understanding, a dark mode for comfortable viewing, and a clipboard copy option for pasting information elsewhere.
Merlin - clicking Cmd+M, you can access ChatGPT on any website - summarize youtube videos
is an amazing ChatGPT extension that works across your entire browser. By clicking Cmd+M, you can access ChatGPT on any website – from Google to Gmail and beyond.
It’s free and privacy-first, and it helps you save time with features like summarizing YouTube videos, generating email responses, and creating AI-driven social engagement on platforms like Twitter and LinkedIn.
is a free and open-source browser extension designed to provide you with an extensive library of prompts, customization options, and additional features.
You can easily discover, share, import, and use the best prompts for ChatGPT while syncing your chat history locally for easy access and searching.
You can export your chats in various formats, such as markdown, HTML, PDF, or PNG
Engage AI - writes LInkedin comments
is a ChatGPT Chrome extension specifically designed to help you write insightful LinkedIn comments, build relationships, and establish connections with prospects. It lets you generate memorable, authentic, insightful, and relevant comments.
Comments need review before publishing
Summary with ChatGPT and Claude is a versatile Chrome extension that uses OpenAI’s AI technology to offer clear and concise summaries of YouTube videos and articles.
It has various features, such as generating bullet point summaries for articles, customization of ChatGPT prompts, or support for both OpenAI API Key and ChatGPT login.
It also offers example prompts, such as summarizing an article in 3 bullet points or 50 words, translating text to Spanish, simplifying text for a 5-year-old, and writing a caption based on the content.
How to Summarise YouTube Videos on ChatGPT (How to Summarize a YouTube Video with Chat GPT)
https://www.youtube.com/watch?v=QrHizi-CSME
Talk-to-ChatGPT
is perfect for when you don’t want to type your queries. It allows you to talk to ChatGPT using your microphone, converts speech to text, and delivers AI-generated responses audibly.
The official extension supports all major languages and is a breeze to use on the ChatGPT homepage.
It’s a fantastic way to simplify your ChatGPT experience, as it utilizes Google Chrome’s speech recognition and text-to-speech technologies.
ReaderGPT
can be a lifesaver for managing your online reading. It generates summaries of any text-based webpage using ChatGPT to help you quickly get the gist of long articles or blog posts.
Its vast array of example prompts in the settings allows you to find the best way to summarize the content you are interested in.
AWS Generative AI Example - Image to Speech
A great #generativeAI example of how to orchestrate multiple #AWS services for an original image-to-speech solution.
Amazon is touting the fact that Q was trained on 17 years of AWS data as a key differentiator for the platform, which is currently available in preview. In one example, Amazon explains how you can use Q to take information from a corporate blog and turn it into a social media post complete with hashtags. In another example, the company explains how you can use Q to create helpful apps for workers.
The service is similar to what Microsoft and Google currently offer across their generative AI platforms. Microsoft’s Copilot and Azure AI studio help workers summarize content and craft emails, as well as create custom AI copilots. Copilots are essentially AI-powered apps.
Google AI pricing
Google’s Duet for Workspaces, meanwhile, acts as an generative AI helper for brainstorming and summarizing content, while its Vertex AI services helps enterprise customers build generative AI apps
Duet pricing = Microsoft 365 AI pricing = $30 per user per month
Vertex AI pricing
https://cloud.google.com/vertex-ai/pricing
Step-by-step guide for Example
sample code block