Christopher Patnoe

https://tr.ee/PfR6UCg_dn SDO - PHX - BOS - New album today. Listen anywhere you listen to music. This is the second of our 3 album commitment  The album "Code to Dust" by The Layoffs explores the theme of transformation and decay in the modern digital age. It delves into how technological advancements and societal shifts affect human experiences, relationships, and identities. The album takes listeners on a journey through various stages of this transformation, highlighting the tension between progress and obsolescence. With a narrative arc that moves from the initial excitement of innovation to the eventual disillusionment and disintegration, "Code to Dust" captures the complexities and consequences of living in a rapidly evolving world.

10

VEO 2 and Google just blew everyone else out of the water. 🔈ON I just came out of a very short night as I just couldn't stop creating and here's the result and a walkthrough: It felt like exploring new worlds, easily blending styles and aesthetics but also being absolutely amazing at understanding small variations. I love my scene but I want snow blowing towards the camera? No problem. I like the video but I would have wanted some claymation? Sure, got you covered. The vast majority of what I threw its way was understood and worked out: night shot, flash light, stop-motion, first person, watercolor, pixel art, etc. I even tried complicated prompts using mirrors and having a person repeatedly reflect in a "bug-in-the matrix-ouch-my-brain" fashion and it worked..you can actually see it in the montage with the man in the blue shirt. Impressive. I have been trying each and every GenAI video tool over the last couple of years and the results for a mere text-to-video are simply in a totally different category than all the others on the market right now...including Sora. I am not even sure where to start: - Resolution in 720p - Prompt adherence is of the highest level, it understands complex scenes and creations, very long prompts with highly detailed explanations of the expectation, camera movements, descriptions, mood, etc. - Format is 8 sec but during those 8 second it understands when you ask for jump cuts or some kind of in-clip editing thanks for the extra angle and scene + character adherence on those cuts VEO. - It is great at creating different styles. It struggled a little with some very specific elements such as asking for an animated oil painting but this is fair... - Generations are fast but it might be because this is an early access - This is the first time I precisely describe the scene I have in my head and actually get it directly or close to it. I had to go for some reroll on some of them but it wasn't because of hallucinations but rather because I felt like I could prompt a little better and fine-tune the output - It is great at filling the gaps in terms of environmental animation - Movements are realistic, some small hallucinations if I try to get a swordfight for example but still lightyears from all the other tools Thanks for the early access Matthieu Lorrain and Google. This came out of nowhere and has been by very far my most satisfying text-to-video experience of the year

440

46 Comments

Never underestimate the power of a small, focused team—they create the space where true innovation thrives. In the early days of game development, small teams were like tightly knit bands of creators, each member playing multiple instruments, harmonizing ideas across disciplines. Conversations flowed freely, leading to raw and organic innovation. The creative energy was immediate—no layers of approval, no rigid hierarchies. As game studios scaled up, the magic often diluted, lost amid processes, production cycles, and sprawling teams. Yet, the charm of simplicity remains powerful, as seen in small, purposeful projects that capture the essence of creativity with just a few hands on deck. Sometimes, fewer voices lead to a clearer vision.

5

Speaking of hallucinations... First, and importantly: I rarely see unexpected hallucinations in my work. I do occasionally see them (although not as much as you would think - certainly not at the percentages being thrown around by those doomer studies), but I only really care when they pop up in situations where I'm not expecting them, and that is pretty rare. Part of this is related to the types of tasks I generally work on. I'm definitely not throwing broad, open-ended legal research tasks at the models. It's just not relevant to my work - and that's intentional. Really, the first step in drastically reducing LLM hallucinations is carefully choosing the right task in the first place. And with the types of tasks that I work on, where I see hallucinations is generally in the early stages of workflow design, when I'm basically just getting a feel for how the models are going to react to a particular task. At this stage, I'm often intentionally trying to get broader, more open-ended responses, to get an intuitive sense of whether the task is possible at all, and maybe an early sense of what categories of errors I'm going to have to solve for. At this stage, hallucinations don't really matter. I am intentionally scoping the task broadly to find the edges, and I know that as the scope broadens and the models are doing more free-form text generation, they are going to occasionally make things up. The key point, though, is that this is just the first step: Once I'm convinced the task is worth pursuing, that's when the real work starts. That's when I can dig in on prompting, testing for the right model, guiding the output format, working through multi-shot workflows, etc. And once all that hard work is done, for me it's pretty rare to see a pure hallucination - meaning, an instance where the models just completely make something up and include it in their response. That's not to say I never see errors or inaccuracies, of course. These are (unfortunately) still not uncommon ("lost in the middle" errors, for example, are much more common than hallucinations). I also sometimes see what I call "quasi-hallucinations" - responses that are clearly grounded in the data I provided but are just weird and not really anything like a human would ever produce. I can talk more about the errors that I do see, but for now, the point I want to make is that, assuming you pick your tasks carefully and then put in the work on the prompt/workflow side, it is very possible to get great results out of LLMs for legal tasks without running into huge problems with hallucinations. The doomer studies you see making headlines with large hallucination rates are getting those results for a reason: they both pick the absolute worst task for an LLM (open-ended legal research questions) and then put no work into prompting or workflow design to try and get accurate results.

3

My first social post ever - So unless #Apple does it, and calls it #AppleIntelligence, any other competitive designer who may have contributed in #onDevice, would've presented, written, reviewed and granted as an #A1Patent 5 years ago, presents it today right before #WWDC as humble part of their contribution in such a scaled & secure 'behavioural design driven' systems feature gets a lash by listening : - 'it was in 2018, why are you presenting it today', - 'what's the point of design when there's no B2C user pain point being solved', - 'we do not understand a clear problem statement, what was the objective' In humility, mind this everyone, #design as an #education and as a practice is as difficult as performing a #surgery, as getting braking as stopping an #aircraft from landing, and as complicated as the craft of saving lives from human error as #Chernobyl. I feel ashamed to be a part of such a community where humility, perseverance towards #planet is understood in the most wrong, unthinkable and extremely downtroden ways as possible. #Shame on such #designers who post a #social influential posts on design #knowhow every hour and call themselves as design managers. This is when the advent of #AI today has lended me about hundreds of in-person discussions, where I convinced my dear parties that there will always be a need to upskill because #behindTheScene, there are people like myself helping #technology reach audience in simplest #withoutUI forms, over a period of time, understanding the consumer and producing longer term learnt - valuable results. Why, to help them in making time for 'criticality'. #Criticality is 'the change' and it's inevitable. But please don't mind, there will be people like me who'll probably kill your job, utterly soon I suppose. That's a challenge accepted in the name of #sustainable future, long long long time ago! Closing this, 'inventions by design' don't propagate beyond single line draft, if they're not based on 'unbiased primary #research' - ability to omit 'your-self'. My audience reading, please look for good #qualitative research rooted designers. You'd not find many, who can 'judge - generalise' extremely pragmatically but if you find one, keep them closer, they'd one day be able to contribute in helping improve your life/lifestyle. Cheers to shared #shame everyone! PS : No more social posts in this decade. I wish we could go back to newspapers and radio. There was one time, luckily, when we had an assignment where we were to #envisage how ( individual #submission) , our #obituary may read. How a future #architect would want it to be for themselves. That was some conversation I had the #opportunity and #privilege to participate and be a part of. Good #education is a gift of a lifetime - this is not subjective. I want #internet + social - everything, to shut, mum for a couple of decades, just shush b, please! Solar storm, please do something. Adios #Linkedin, the new #Facebook.

15

Here’s a super clear and practical blog post from Jason Myers that explains why and how we lint our code at #Klaviyo. https://lnkd.in/eeTSJJD2

3

As someone who relies on their Fitbit (now part of Google) watch for daily insights, I'm struggling with the recent app redesign. The previous interface was intuitive and user-friendly, but this new design seems less accessible and difficult to navigate. It's a reminder that innovation can sometimes detract from user experience. Can anyone share insights into the rationale behind these changes? #Fitbit #UserExperience #Design #Innovation

12

6 Comments

🍩🍩Take a look at Campaign Donut 🍩🍩 A free, micro unmanaged package for Salesforce Campaigns. Salesforce got rid of the Doughnut Chart at the top of the Campaign list view, Report Charts are slow and cache results aggressively. Meet Campaign Donut, a Lightning Web Component that: 🍩 Goes on the Campaign Page as a doughnut graph 🍩 Refreshes when the record is opened 🍩 Navigates to the Campaign Member list view when clicked I also added in a table with the total and the breakdown by status. Thanks to this (https://lnkd.in/eaMHE4BS) post from Jodie Miners, as my motivation for a holiday weekend project (plus users in my org have been asking about it too). App Exchange Links: Production/DevEdition: https://lnkd.in/eNviaW3Y Sandbox: https://lnkd.in/eaaaYWdt Installation: 🍩 Install Unmanaged Package 🍩 Add sb_campaigndonut to lighting page layout (don't know why but it took about 5 mins to appear as an option) 🍩 Either give the packaged Permission Set (Campaign Donut) or assign access to the class (sb_CampaignDonutController) Components: 🍩Lightning Web Component (sb_campaigndonut) 🍩Apex Class (sb_CampaignDonutController) 🍩Static Resource (chart.js) 🍩Permission Set (Campaign Donut) Let me know what you think! #salesforce #marketingOps #salesforceAdmin #ohana #marketo

21

2 Comments

I've spent a lot of time recently setting up teams for collaboration, along with meeting new people and figuring out how to work with them. As I reflected, there's something important that came up, not only for collaboration, but any interaction. Context. If someone expresses an opinion you don't agree with, it might be because they have a different context. They took a different path through life, with different experiences, to land on that opinion. It's true of every opinion a person might have. Programming languages, frameworks, politics, the weather. The personal context will influence what they're passionate about. It's a weird reflection in modern society that it feels like everyone must care about everything equally all the time. But people usually have coding, causes, and features they're more passionate about based on their context. Understanding this, and sharing with colleagues, is crucial for collaboration to be successful. It's way easier to have a disagreement or debate if you understand the other people involved, and the context they're coming from. https://lnkd.in/dkzwW9Dz

6

My brain is expanding rapidly. I’ve had an incredible first few days on my Snap Lens Lab programme with Snap Inc. and Royal College of Art. Yesterday was spent at the Snap office where we were able to learn so much about: - The culture at Snap and how employee resource groups are empowered, championed and a core aspect of the working environment. - Experience AR technologies IRL with the AR glasses which was very surreal and from my technical perspective just ignited so much curiosity and interest even if I didn’t understand all the physics of it 😂. - Hear unique perspectives around branding, communications, creative strategy and product. So much I could say but I want to focus on my gratitude for this opportunity which I can already feel is having such an impact just by being able to connect with other like-minded creatives and absorbing a tonne of knowledge from employees/ creatives at Snap and the RCA. One key point I took from Valentina Culatti that was echoed by others was around letting go of your ego and being humble which is a value that keeps me grounded on a personal level but one that I believe is fundamental for the success and evolution of industries and individuals. Cant wait to see how my knowledge of immersive technologies deepens and what can come from this experience. #womenintech #augmentedreality #tech

69

6 Comments

Recent article in Raconteur about how we've been using GenAI to improve accessibility https://lnkd.in/eTnzxANa

13

1 Comment

I'm a fan of niche devices, but many struggle to gain traction. Here's my take on the Tiny Pod. I am somewhat of a fan of niche devices -- single, or less multi, purpose and by design less complicated than what our smartphones have become for example. Yet, most such devices have failed to get enough traction -- I mean they are niche by choice so can't expect them to become a mass market phenomenon, but at least they could find a niche on which they can thrive on? Some devices I have bought and have gotten off myself after limited usage: Spotify Thing, Play.date, Google Cardboard, Pebble Watch, and Misfit Shine. I currently own and do use an Oura Ring (I like it so far, but let's see if that continues over the long run), and may give Play.date another chance. This brings me to this really cute and smartly thought device: https://thetinypod.com/ -- the concept is so cute and price low enough that I may just buy it to play with. However I can't see myself using it over the long run because of the following reasons: - I don't think I'd want to give up on wearing my Apple Watch on my wrist; I like the idea of going out without my phone but do like to keep my watch on the wrist. Using Tiny Pod thus would mean buying another Apple Watch (and a cellular plan!) with cellular connectivity. - I mostly use my Apple Watch for consuming stuff or as a fitness tracker -- for example I have rarely if ever used it to send messages, but of course use it often to read them as well listen to music etc. -- activities that do not (for the most part) require me to create or input stuff. Therefore this is where Tiny Pod falls short; it is limited to the same interface and capabilities that Apple Watch provides. - As far as I know I don't see a Whatsapp app for Apple Watch -- I can get notifications but since there is no watch app, I can't send messages. It boggles my mind that most people in the US live on Messages (iMessage, or any form of SMS) in this day and age; it's practically impractical to do so for those living outside the US or those with a critical mass of their contacts on Whatsapp (or another similar platform). PS: I'd like to quibble about the available colors on the Tiny Pod -- since one primary reason for such devices is their cuteness appeal, would have been nice to have some colorful options. PS 2: I have written extensively about how smartphones have become complicated and why we need "digital divergence" of sort here: Smartphones Do Too Much: Convergence is Giving Way to Divergence: - https://lnkd.in/grdZG9aV - https://bit.ly/4btYPR2 - https://bit.ly/3W7uDa9 How AI can simplify the smartphones and what are the prospects of some emerging platforms (Rabbit's R1, Humane's Ai Pin) trying to do the same: - https://bit.ly/48CCyQv - https://bit.ly/3vW8PUe #TinyPod #DigitalDivergence #NicheDevices #Wearables #Smartwatches #SimplifyingTech #TechReview #AIinTech

1

Not sure about this implementation but this is the direction things are heading for sure. If its not the Humane Pin, it'll be Friend, if not Friend, something else. We will likely not stop anthropomorphizing tech until there is a huge paradigm shift. The question of trust and usefulness comes immediately to mind... but as these elements become smaller, open source and own-able, trust has a better possibility of being baked in. Usefulness...the devil's in the details. https://lnkd.in/e5KSPGn2

7

1 Comment

Back in September, I watched Meta's reveal of the prototype for the Orion VR Glasses. Instantly, I thought, "They're going to make another smartphone!" Read the latest post to learn more about how I made this connection and why I think this would be a smart strategic move for Meta. #Facebook #Meta #VR #AR #AI #FutureOfTech #FuturePhone https://lnkd.in/gUaRFmB6

7

UK Filmmaker and YouTuber, DSLR Guide, shares how he collaborated with an editor remotely on his latest project, using Dropbox Replay - a collaboration tool that allows users to fast track feedback on their audio, video or image content, no matter where they're based. Watch below to see how DSLR Guide uses Replay to reach that final version faster... 👇

2

1 Comment

Is Keir Starmer Hoping AI is the Silver Bullet for the UK Economy? On January 13, 2025, Keir Starmer appeared on live television to unveil a bold new vision for the UK: full-scale adoption of artificial intelligence (AI) across both public and private sectors. With an impressive £14 billion in commitments from leading corporations, this initiative—dubbed the AI Opportunities Action Plan—aims to position Britain as a global leader in the AI revolution. However, when compared to nations like the United States, China, and Saudi Arabia, which have each pledged hundreds of billions towards AI development, the UK’s investment appears modest. This disparity underscores the challenge of competing on the global stage and highlights the need for strategic allocation of resources to maximize impact. https://lnkd.in/eFzYGxE9 #AIRevolution #UKEconomy #ArtificialIntelligence #AIJobs #KeirStarmer #AIIntegration #FutureOfWork #DigitalTransformation #DataPrivacy #Reskilling #AIOpportunities #AIChallenges #TechPolicy #AutomationImpact #GlobalAICompetition

3

1 Comment