Alex Banks Profile picture
Apr 27 11 tweets 4 min read Twitter logo Read on Twitter
Yesterday Palantir announced its Artificial Intelligence Platform.

Here's how it transforms the future of military and defence: Image
AIP has 3 core pillars:

• Deploys LLMs and AI across classified networks & devices
• Connects classified data to create a real-time environment
• Enables security features controlling AI access

I feel guardrails & security are critical for sensitive data here.

Let's dive in Image
We'll start with an example: you're a military operator.

You get an alert of activity nearby & ask AIP to show more details.

AIP:

• Uses satellite data & AI to build likely battle formations
• Surfaces the idea of deploying a nearby drone to collect video
You task the drone to capture video.

The footage confirms a threat and shows a tank.

You ask AIP what to do next.

AIP:

• Generates 3 courses of action to target the enemy equipment
• Automatically sends these options up the chain of command
Let's take a look at what's happening behind the scenes.

As you ask questions, AIP:

• Traverses information across both public and classified sources
• Enforces which part of the organisation the LLM has access to
• Auto tag & protect data respecting permissions
You now assess the courses of action generated by AIP.

AIP:

• Highlights key information
• Find personnel & armament requirements
• Determines time required & status of supplies

It then establishes the closest unit before building an automated terrain analysis model.
Things get even crazier.

AIP:

• Suggests optimal route based on terrain analysis
• Identifies and disrupts enemy communication nodes

You can then use AIP to summarise and submit the operation plan.

AIP has successfully helped you determine the course of action.
I'm blown away by this application of AI in military situations.

I can only feel it makes AIP a critical tool for modern military operations moving forward.

The takeaway?

AI-powered militaries are quickly becoming a reality.
Follow me @thealexbanks for more on AI.

If you liked this thread, you'll love the newsletter.

Cut through the noise in AI each week, subscribe here:
noise.beehiiv.com/subscribe
Help everyone learn and retweet this thread:
You can read more about it in Palantir's announcement here:
palantir.com/platforms/aip/

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Alex Banks

Alex Banks Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @thealexbanks

Apr 25
Long-term memory is coming to Bing Chat.

Here's how it will revolutionise the future of AI: Image
Mikhail Parakhin is the CEO of Advertising & Web Services at Microsoft.

He teased that Bing is soon incorporating memory in a fairly restricted form.

This opens up a world of opportunity for the platform. Image
1/ Long-term memory feature for Bing Chat

Bing will be able to:

• Retain information from previous chats
• Improve context awareness in ongoing conversations
• Enhance user experience with personalised responses

I feel this will make Bing more intuitive and user-friendly.
Read 8 tweets
Apr 23
I came across the human progress curve from Tim Urban in 2015.

But it took me until yesterday to realise that:

• 5 months ago: ChatGPT was released
• 5 weeks ago: GPT-4 + plugins were released

Here's why AI is accelerating far quicker than you'd ever think: Image
Ray Kurzweil is an American computer scientist and futurist.

He believes that the 21st century will achieve 1,000x the progress of the 20th century due to a single driving force.

The Law of Accelerating Returns.

This suggests that human progress accelerates as time progresses.
In simple terms: societies with advanced technology can improve faster than those with less advanced technology.

By 2000 the rate of progress was 5x faster than the average rate of progress during the 20th century.

But what about Moore's Law?
Read 11 tweets
Apr 20
A new era of AI content creation has begun.

Nvidia just announced its text-to-video model.

Here's how it's going to transform the future of video forever: Image
1/ Nvidia releases Video Latent Diffusion Models

Let's start with the basics.

A Latent Diffusion Model (LDM) is a way of creating realistic images or videos using AI.

It learns patterns from other images without needing a super-powerful computer.
2/ Text-to-video synthesis with Stable Diffusion

Nvidia leverages Stable Diffusion for text-to-video generation:

• Resolution up to 1280 x 2048
• 4.7-second-long clips at 24 fps
• 4.1B parameters

Nvidia adds a time-related aspect to the model to turn it from image → video.
Read 8 tweets
Apr 18
Bloomberg just introduced BloombergGPT.

It's a 50-billion-parameter large language model for finance.

Here's how traditional finance is quickly getting disrupted: Image
1/ BloombergGPT's training

BloombergGPT is trained on press releases, news articles and filings.

• Half domain-specific text, half general-purpose text
• 363 billion token dataset from Bloomberg's archive
• 345 billion token public dataset

Now let's explore some use cases. Image
2/ Use case: Streamlined research

BloombergGPT helps make interactions with financial data more natural.

Current data retrieval is done using Bloomberg Query Language (BQL).

This is a powerful but complex tool.

BloombergGPT can transform natural language queries into BQL. Image
Read 14 tweets
Apr 16
I read all 154 pages of Microsoft Research's "Sparks of AGI".

Most people don't realise the shocking reality AI is going to bring.

Here's what I found:
1/ Tool use

GPT-4 suffers from serious limitations:

• Lack of world knowledge after September 2021
• Difficulty with math
• Can't execute code

However GPT-4 can use external tools like search engines, APIs and plugins to overcome these limits. Image
For humans, the use of tools was a milestone in history.

It allowed us to overcome the limitations of our physical abilities.

The same can be said for GPT-4.

It can use tools with:

• No demonstration
• Minimal instruction
• Make appropriate use of the output Image
Read 14 tweets
Apr 13
Stanford just released a groundbreaking AI study.

"Generative agents" can simulate authentic human behaviour.

Here's the breakdown (and the potentially devastating consequences): Image
1/ Setup

Smallville is a Sims-style game world.

There are 25 generative agents (characters) who live in the game.

They operate in a way that closely resembles human behaviour.

Agents can:

• Plan their day
• Share any news
• Form relationships
• Coordinate activities Image
2/ Structure

How do you make generative agents?

You need an architecture that stores, synthesises, and applies relevant memories to generate believable behaviour.

This is done using a large language model—ChatGPT.
Read 13 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!

:(