#cdnpoli - Feds sign another MOU with another Canadian AI company under the Shared Services Canada department. There's no budget transparency here in terms of how much they'll be spending on these 3rd party contracts.
An evergreen reminder that you are not your users. Go talk to people who use your stuff, or might want to use your stuff and see what they really want.
I wonder how much of our writing style is going to shift because of the massive unconscious exposure to slop.
May's tech-hype predictions for 2026 onwards: - The month after major public pensions and governments buy huge stakes in the AI companies (OpenAI, Anthropic, etc), the US AI bubble will start crashing. This will happen in 2026. - At this point, free access to LLMs will go away. Corporate LLMs will raise their prices by at least 100x to make up for the loss in users and casual revenue. - Unfortunately, this^ means AI coding /marketing wouldn't go away. It'd just be far, far more expensive and require deeper approvals. Your boss will have AI and you won't be able to summarize it. - China and the rest of BRICS will take over the free AI market because they can. Perfect opportunity for propaganda and cyber warfare here. - Many software products that wrap base commercial models will see their operating costs skyrocket. They'll basically need to rip it all out, or embed a cheaper model. - Essentially, companies who haven't built out a proper data/AI team will fail unless their base non-AI product is still relevant. - Small companies will die when they fail to transition out of this mess. There will be a wholesale buyout, like the dot com crash. This means deeper consolidation at the FAANG level. - We've had essentially an entire generation of juniors who had a cheap/free magical crutch for years. How will you train people to read, write and code again across every industry? - The costs of cloud services will also skyrocket, since they need to pay off their bullshit AI hardware/data centre investments. We may see a return to on-prem/hybrid IT. (This will be REALLY interesting for how software evolves from the current SaaS model.) - Some people will start hyping quantum, but we're at least 2+ hype cycles away from that. - (This one's a wild guess.) The next hype cycle is probably going to be in hardware/wearables/physical space, but that's ridiculously expensive to get into. It'll basically just be marketing. - The techbro doomers will claim to have predicted this the whole time. They'll have more than enough money to start the next hype cycle for the next round of VC/PE hot potato. - Western governments are still going to try to implement "AI" even after the crash and the prices go up.
You ordered a pizza, cash on delivery. 35 minutes later, a cop car pulls up to your door. The cop knocks on your door. The cop is holding a pizza. What do you do? #poll
I need you to stop scrolling and read the most human piece of writing I've read in a very, very long time. Thanks, @npub159p8...dhu3.
I hope this 3-hour planning meeting invite without an agenda never finds you.
Today's gender: Joybringer with a tinge of sleepiness.
I have a love-hate relationship with my inner monologue. More like rambles. These go on forever! What ever happened to being succinct?