Weekly Link Roundup #23
With lots of stories about a certain AI company doing stuff they should've done before launch. Oh yeah, and AI is already in your orgs....
Microsoft: Workers want to use AI—they’re not waiting for their companies to adopt it by Ken Yeung: First, Ken is a brilliant journalist and you read all his stuff and hire him and second, TOLD YA SO. If you’re still dithering about wondering what the right AI strategy is for your org, you might want top ask your employees since they’re already using it. Much like they are already learning without courses from your LMS. See also: Burnout Is Pushing Workers to Use AI—Even if Their Boss Doesn’t Know,
Forget about prompt engineering, Typeface Arc uses AI agent approach to power marketing campaigns: Now look, if you can read the following quote and not see how this could move into creating, oh I don’t know, say training content, then I don’t know what to tell you - “Parasnis argued that many of today’s gen AI systems are limited to disjointed outputs. While generating a single new piece of content provides moderate usefulness, the ability to choreograph personalized customer journeys from start to finish is still needed.”
Microsoft is ‘turning everyone into a prompt engineer’ with new Copilot AI features: Why does reading this make me think of Ouroboros and this image? So AI will suggest
ways for you to write a better prompt………for AI.
Dell cracks down on hybrid working again — computing giant is going to start color-coding employees to show who is coming back to the office: The definition of a corporate face palm. Also the definition of a failure of imagination and innovation by management. Seriously, thought I’d seen some dumb RTO orders but the color coding here is just chef’s kiss idiocy. Also just made my decision easier to not buy an Alienware laptop.
OpenAI posts Model Spec revealing how it wants AI to behave: Just like standing up a safety group AFTER LAUNCH, this again feels like something that would have been good to work on and send out BEFORE YOU RELEASE YOUR PRODUCT. Sheesh. Not its just cheap CYA. And this bit from Altman “we will listen, debate, and adapt this over time, but i think it will be very useful to be clear when something is a bug vs. a decision”… just defines wishy washy…have a take, stick to your position, don’t say upfront that every time you get confronted with something, you’ll change your mind. Gosh, what could Model Spec be providing cover for? Oh I don’t know.. OpenAI plans to announce Google search competitor on Monday, sources say. Or gosh, maybe this: Leaked Deck Reveals How OpenAI Is Pitching Publisher Partnerships. It just keeps going: OpenAI considers allowing users to create AI-generated pornography. Judas….here’s another one that could’ve saved them so much trouble, OpenAI says it’s building a tool to let content creators ‘opt out’ of AI training.
Billions in Chips Grants Are Expected to Fuel Industry Growth, Report Finds - The United States will triple its domestic chip manufacturing capacity by 2032, the largest increase in the world, according to a report from the Semiconductor Industry Association: Not as important as access to clean water but certainly key going forward.
Sakana AI Releases Japanese DALLE-3, Calls it EvoSDXL-JP: We should be adding multi-cultural to multi-modal to our considerations on what LLMs are trained on.
Teens and Video Games Today (Pew Report): “85% of U.S. teens report playing video games, and 41% say they play them at least once a day. Four-in-ten identify as a gamer. 72% of teens who play video games say that a reason why they play them is to spend time with others. And some have even made a friend online from playing them – 47% of teen video game players say they’ve done this. Over half of teens who play video games say it has helped their problem-solving skills, but 41% also say it has hurt their sleep. 80% of all teens think harassment over video games is a problem for people their age. And 41% of those who play them say they’ve been called an offensive name when playing.”
Researchers at NVIDIA AI Introduce ‘VILA’: A Vision Language Model that can Reason Among Multiple Images, Learn in Context, and Even Understand Videos: As Ethan Mollick always reminds us, the AI you’re using today is the worst AI you’ll ever use. The kind of leaps forward like those in this article make two things clear: more than ever, your org needs to direct some % of its work to staying current on developments (hello opportunity for L&D) and you need to be engaging with some foresight work so that you can envision some of the future outcomes that are possible.
Bluesky to add DMs, video support and in-app custom feed curation: Define too little, too late. Look, Twitter could work when it launched because there was no competitor. Now we have Threads, Mastodon, Substack is launching a similar feature set, and on and on.
I regret to inform you that LinkedIn’s games are very fun: I get why they’re doing this - Wordle - but sheesh, talk about moving away from your core product and probably your core audience.
Atlan scores $105M for its data control plane, as LLMs boost importance of data: I keep saying it, but in a gold rush, look for the people selling picks and shovels. This kind of capability can be the last mile work in terms of being able to use all that data your org has been collecting for years.
Sperm whale ‘alphabet’ discovered, thanks to machine learning: IFYKYK “Calm down, get a grip now … oh! this is an interesting sensation, what is it? It’s a sort of … yawning, tingling sensation in my … my … well I suppose I’d better start finding names for things if I want to make any headway in what for the sake of what I shall call an argument I shall call the world, so let’s call it my stomach.”
Did Stanford just prototype the future of AR glasses?: No notes - just here for it.
Making Third Places First in Your Heart - Neutral territories that are not home and not the workplace should be a high priority for urban planning: Yes, yes, a thousand times less. We just have to find the funding model/political will to do this.
Russo Brothers’ AGBO launches innovation department for tech creativity: This should be interesting “the AGBO Innovation department will focus on developing proprietary creative tools for storytellers, fostering production innovation, and enhancing cost efficiency across diverse platforms.”
Stack Overflow is feeding programmers’ answers to AI, whether they like it or not: The news just seems chock full this week of orgs that are shooting themselves in the foot. Hey, Stack Overflow, show me how you can really piss off the people who generate all the content on your site - “Users have been barred from deleting answers, and Stack Overflow moderators have restored answers that were changed in protest of its new deal with OpenAI.”
Sprinklr launches ‘Digital Twins,’ AI versions of brands to provide better customer experiences: Told ya so. Again, told ya so. The AI-Generated Population Is Here, and They’re Ready to Work.
OpenAI Says It Can Now Detect Images Spawned by Its Software—Most of the Time: Or and hear me out here, you could just add a feature to your own image generator that adds machine-readable watermarks to its images. But that’s just me.
Newsletter platform Ghost adopts ActivityPub to ‘bring back the open web’: This makes me hopeful that we can still step back from all the walled gardens.