62. Stop Learning AI Prompts, and Apple’s Siri Fumble.

Hi all. Two things caught my eye that would compel me to share. First is Prompt Buddy, that is a must read. Secondly we have Apple’s Siri/AI FUBAR. The third one was about a holding company purchase that I cut. It seemed to dilute the other two articles and was a bit blah. So you’ll never know about it.

Shall we?

I get over-excited about most things that are smart, innovative, and can be used immediately. How ironic it is that to activate AI, you’re required to learn or memorize countless prompts, and manually type them in. Loads of data entry.

But what if AI could elaborate and re-write your prompts to make your imagery more detailed. More colorful. More real.

It can.

I’ve seen a couple of these floating around but this so far is my favorite.

The Veo 2 Prompt Buddy is a custom GPT tool developed by Allar Haltsonen to assist creators in generating cinematic prompts for Google DeepMind’s Veo 2 AI video model. This tool is also compatible with other AI video platforms like Minimax and Hailuo.​

Veo2 Prompt Buddy will take whatever scene description and turn it into fascinating, detailed, and compelling motion footage. It augments your original prompt with additional, relevant descriptors that add complexity and depth to your production.

While I don’t have a case study to present, I can share a compelling before/after scenario.

Frame grab from a dolly shot left to right – without Prompt Buddy.

Frame grab from a dolly shot left to right – with Prompt Buddy.

The Prompt Buddy offers four distinct output modes to cater to various creative needs:​

1. PRO: Delivers detailed, professional, and cinematically accurate prompts.

2. OUT OF THE BOX: Generates creative and unexpected ideas.​

3. ULTRA DETAILED: Expands on ideas with extreme precision and detail.​

4. CAMERA COMBINATIONS: Provides cinematic camera transitions inspired by the Hailuo I2V-01-Director model.

Regarding the above simple prompt example, here’s what you get.

You select the prompt of choice, copy/paste into your video generator of choice, and away you go. I’ve only used it with Sora so try it with motion or even image generators like Midjourney or ChatGPT (minus camera moves, etc.)

Ready to begin? Bookmark this link. Veo2 Prompt Buddy

Share

Apple’s ambitious plans to revamp Siri with advanced AI capabilities have encountered significant setbacks, leading to delays and internal challenges. Here’s an overview of the key issues:​

Apple’s AI and machine learning division, led by John Giannandrea, faced criticism for indecisiveness and a lack of clear direction. Engineers were frustrated by frequent shifts in strategy, such as oscillating between developing multiple language models and a single cloud-based model. This inconsistency led to project delays and staff departures. Internally, the AI/ML group was nicknamed “AIMLess,” reflecting perceptions of its unfocused approach. MacRumors​

The company officially acknowledged a delay in a statement to Daring Fireball, stating that the features would roll out “in the coming year.”

“Siri helps our users find what they need and get things done quickly, and in just the past six months, we’ve made Siri more conversational, introduced new features like type to Siri and product knowledge, and added an integration with ChatGPT. We’ve also been working on a more personalized Siri, giving it more awareness of your personal context, as well as the ability to take action for you within and across your apps. It’s going to take us longer than we thought to deliver on these features and we anticipate rolling them out in the coming year.” Tom’s Guide

Efforts to enhance Siri’s AI capabilities were hampered by budget cuts and outdated hardware. Despite initial approval to double the AI chip budget, CFO Luca Maestri reduced the increase, resulting in reliance on less advanced GPUs. This decision delayed development and limited the performance of AI models. LaptopMag

Share The Ad Stack

At WWDC 2024, Apple showcased Siri features like contextual awareness and multi-app task execution. However, many of these capabilities were not fully developed at the time. The demonstration surprised some Siri team members, as the features presented were not yet functional on test devices. MacRumors​

Robby Walker, the head of Apple’s Siri team, referred to the delays and miscues as “ugly” because the overhaul has already received heavy public promotion, including in TV commercials, Bloomberg reported, citing employees with knowledge of his remarks.

The development of Siri’s AI features faced technical hurdles, including bugs and inconsistent performance. Some AI functionalities generated incorrect responses up to one-third of the time. These issues led to further delays, with the full rollout of enhanced Siri features now expected with iOS 19 in Fall 2025. ​New York Post, Tom’s Guide

Apple’s strong emphasis on user privacy influenced its AI development approach, favoring on-device processing over cloud-based solutions. While this aligns with Apple’s privacy commitments, it posed challenges in implementing advanced AI features that competitors achieve through cloud computing. ​

In response to these challenges, Apple reassigned leadership roles, with Mike Rockwell taking over Siri’s development. Craig Federighi’s software team, known for delivering results, has been granted more flexibility, including the use of open-source models. These changes aim to reinvigorate Siri’s development and align it with Apple’s broader AI initiatives. ​Marketwatch9to5Mac

Despite these setbacks, Apple remains committed to enhancing Siri’s capabilities. The upcoming iOS 19 release is anticipated to introduce significant improvements, positioning Siri to better compete in the evolving AI landscape. However, they look like less of a market leader or a market maker and more of a leading product alternative. I doubt this will affect share amongst most Apple diehards including myself. My iPhone 11 Pro Max can hang in there for another year.