192 | 🚨🚨BREAKING: Elon Musk to develop X into AI news site

In today’s Brainyacts:

1. OpenAI’s new webpage is NOT good

2. Anthropic’s Claude goes mobile

3. Palantir solves legal RFP problems and other AI model news

4. Military AI goes nuclear too often and other AI-related content

👋 to new subscribers!

To read previous editions, click here.

Lead Memo

🧑‍🎨👩‍💻 OpenAI updates its homepage and why this is not a good thing.

Rule #1 about prompting conversational generative AI - it is not search!

If you want to watch a short video on this, instead of reading, click below:

As rumors circulate about OpenAI potentially developing a search engine to rival Google, excitement and speculation are growing. The prospect of reinventing search to be more intuitive and helpful is undeniably appealing. I mean who isn’t sick of the first search result being paid ads? So questions about monetization strategies, such as whether OpenAI will adopt an ad-based model, are crucial. But they are secondary to me considering I focus on the pragmatic use and implications of generative AI for knowledge workers, particularly in the legal field.

In my experience teaching the pragmatic use of generative AI to 1000’s (not overstating that) law students and legal practitioners across all stages of their careers, I’ve observed a significant challenge: the transition from traditional search habits to effective interaction with conversational AI. Traditionally, search engines like Google have hard-wired us to input minimal information—often just a few keywords—to receive answers. This method encourages a sort of cognitive laziness, a reliance on the search algorithm’s ability to interpret and respond to limited input.

However, engaging effectively with conversational AI, such as the type developed by OpenAI, requires a different approach. This technology thrives on detailed, thoughtful queries. When users input more comprehensive, context-rich questions, they receive more precise, actionable responses. This distinction is crucial for legal professionals who are accustomed to the nuance and specificity required in legal research and documentation. 

The interface that OpenAI proposes, reminiscent of Google’s minimalist search page, risks perpetuating the old habits that are ill-suited for conversational AI. By presenting a simple input field, it invites users to continue entering minimalistic queries. While the underlying models of generative AI may be advanced enough to handle some level of vagueness, this is not the ideal use case. Insufficient input often leads to generic or less helpful responses, which can quickly sour users’ perceptions of AI’s utility, branding it as inefficient or unreliable.

For legal professionals, the stakes are even higher. The precision of legal advice, the accuracy of legal precedents, and the appropriateness of contractual language cannot be left to the uncertainties of an under-informed AI. Legal queries, in particular, benefit immensely from detailed background information, clear context, and specific questions. The effectiveness of AI in legal applications hinges not just on the technology’s sophistication but on how lawyers engage with it.

To address this challenge, OpenAI and similar entities need to consider user education as part of their rollout strategy (or leave it to those of us who teach it). It’s not enough to build and release an AI-powered search tool; they must also guide their users towards best practices in interacting with this new technology – or at the very least, draw a clear distinction between traditional search engine-style query and generative AI-style query. Lumping them together via a similar web page design is not just unhelpful, but downright destructive to adoption of generative AI. 

Since you are a read of Brainyacts, hopefully, you know all of this. But if you know of someone new to generative AI or skeptical, share this with them so they stand a better chance of not judging their interaction with these tools based on the responses they get – at least until they realize what they are inputting is 10x more important than it ever was before in search.

BTW. The web address search.chatgpt.com looks to be live but empty (for now).

Spotlight

📲📸 Meet Mobile Claude

Anthropic released a mobile version of Claude (iOS only right now). It is free to use but if you want access to 5x more Claude usage and our most powerful model (Claude 3 Opus), you will need to upgrade to their paid plan of $20/mo.

The app features:

• Syncing with web chats: Continue your conversations seamlessly across different devices.

• Vision features: Access and analyze images from your photo library, capture new photos, or upload files for on-the-go image understanding and mobile use.

OpenAI is still my “go-to for” a great mobile experience. But it is great to see this other leading model trying to compete and go at them. It only benefits us, the users.

AI Model Notables

► Palantir is IMO the de-facto leader in enterprise/large organization (e.g. BigLaw) generative AI use. Any lawyer or business professional (e.g. legal operations) out there who creates, manages, or responds to RFPs should watch this! It is incredible and solves many notorious problems like:

• knowledge management

• consistency of brand and service descriptions

• compliance or partner or GC-approved language

• wasted partner time starting from scratch in responding to RFP

• detailing/calculating pricing for different segments of work

• and so much more.

► gpt2/powerful new chatbot disappears as mysteriously as it arrived

► Microsoft bans US police departments from using enterprise AI tool for facial recognition

► Microsoft is looking to build its own humanoid robots

► Microsoft to generate 10.5 gigawatts of renewable energy to power its AI

► Apple to unveil AI-enabled Safari browser alongside new operating systems

► Apple has advantages that will set its generative AI apart from the competition

► Anthropic launched a $30/month Team plan for Claude featuring increased usage access and a 200k context window, alongside a new iOS app for mobile access to the chatbot

News You Can Use:

➭ Why the military can’t trust AI: tldr - they tend to go nuclear quickly PLUS why the US is urging China and Russia to follow suit

➭ Ukraine unveils AI-generated foreign ministry spokeswoman

➭ AI discovers over 27,000 overlooked asteroids in old telescope images

➭ 49 countries and regions have signed up to Japan's international framework for regulation and use of generative AI

➭ AI trained and AI robot dog to balance on a yoga ball

Who is the author, Josh Kubicki?

Some of you know me. Others do not. Here is a short intro. I am a lawyer, entrepreneur, and teacher. I have transformed legal practices and built multi-million dollar businesses. Not a theorist, I am an applied researcher and former Chief Strategy Officer, recognized by Fast Company and Bloomberg Law for my unique work. Through this newsletter, I offer you pragmatic insights into leveraging AI to inform and improve your daily life in legal services.