I ran multiple experiments with models the past few weeks for generating designs and 80% of the time they use Fraunces if no selection is provided. Even if this was not AI generated it will appear as if it is because of this Font. It’s the new em-dash.
Can we all agree to system prompt Claude and open models to stop using the same exact Fraunces font everywhere (The Capital J looks like someone folded it backwards). I said open models because qwen and Glm and Claude produce the same design and font for the same “design a gorgeous website” prompt.
Intelligence and the mental load it requires has a budget - in both real and artificial forms. In real - people are bombarded with tools and requirements and when the teams behind those don’t consider mental load required, that spends down the budget. Thus people give up fighting all battles and are forced to prioritize their resources. Denial of service attack on the brain. AI models running out of context will do the same and get lazy. Sadly any bad actor private or government knows this and uses it as a tool to fight intelligent resistance to its plans.
Because of statistics. If we can find life, potentially started independently or pollinated in a couple of planets close to each-other, then life is everywhere and we may be in a dark forest of intelligent or hidden native predatory beings waiting to wipe us out if we squeak or show up elsewhere. If it's not as frequent, then we can keep exploring to look for it and more human-habitable spots.
It goes from complaining about endless pitches of tech to pitching a gaming console. The tech pitching is a problem, but the need to be constantly entertained and fed some content or game is the real problem. Those are pacifiers. You stuff them in an adults mouth to keep them from actually thinking about what would improve their own life. Take the pacifier out and use your brain for your needs - what you should want to do, to make of the world, to make exist. See what happens next…
Reminds me of a display at Chabot in the bay area. A bubbling volcano covered by a stable mass stays bubbling underneath, but shaking up the mass gives little bubbles a bath up and they break way for the big bubbles leading to a volcanic eruption. If you fragment a he soil at some point in the mantle of earth and there is pressure anywhere below, you could get yourself a new volcano.
In the past few could afford to be photographers due to time and money. Many were not the most artistic or talented minds in the universe on the subject (you can’t find those until you give everyone a camera) but just people who enjoyed it and could make money off of it. Those who gained experience with the old tech were essentially AI models trained on statistically good probabilities of getting good shot. You had smaller selection and the model / brain quality determined what % of
the best that photo was between 0 and 100% for any sunset. In modern days more people get a camera - more people with bad taste and more people with good taste. You also get more attempts at the shot - more photos taken, more tries. More tries means you are now preserving more of that moment from more angles and extending the time your taste people have to pick the best shot of all possible.
Your plethora of statistically average photos happen if you don’t have anyone filtering the ginormously increased output. If the problem is really valuable, someone will filter it. Possibly a model.
The real filter is value, not capability anymore. Do you have the empathy and taste to know what matters? The AI tokens/cost lets everyone make and become interested in making of someone hasn’t solved their problem. And then if search isn’t improved - everyone will just repeat the making of the same thing over and over. So if you don’t want repeated slop. fix search.
Or do you copy everything and everyone into making more of nothing.
You are a photographer from the olden age. Taking a photo back then took forever. You are standing on a hill trying to get the perfect picture of a sunset. You measure and sort and calculate angles and light. A tourist stands next to you with an high-speed iphone auto-settings camera. They spin around and get thousands of photos from different angles while you are thinking and measuring your one shot. That’s what AI does. It gives you many shots at the same problem. The thinking guy with the one approach will likely lose a photo competition to one of the accidentally better shots of the person who tried a lot more. Make and filter now may be a better approach than overthink and try only once.
Are we considering what is the shelf life of information?
If you build a building, the expense on materials determines longevity. If you build a city. The robustness of government and the economy in it determines the property taxes and value of property over time.
If you make or cook food. The majority of the nutritional value of it goes to the initial consumption. Once the food has stayed out without refrigeration it is taken over by bacteria and fungi. Refrigeration seems to be paywalls. Once the information is out it accumulates at exponential rates - the amount of text on the internet does not diminish but increases. Some people may “prune” old content away, but that is rare. Human attention is somewhat a fixed number. Thus text left out is not consumed, but sits idle and decays in accuracy and value over time. The fresh content of valuable should be in a fridge. If not valuable it is released - thus scavengers and those hungry and motivated to dig can consume it. If spammy and sales-y / propaganda-y which a lot of content farms are doing, the goal is for it to be consumed by the masses and push the zeitgeist to buy its premise. That’s Sugar or addictive shelf-stable junk foods. AI model companies are the bacteria / fungus/cockroaches/rats of the information dumpster. They sneak out any remaining energy from content that would otherwise be buried by other content and try to give it a second shelf life - one reachable and accessible and consumable by humans. They make alcohol. Alcohol is addictive. Ir may mess with your brain - it may make you lazy. It will sneak in bad decisions because it lowers your judgement. It is repurposed food, not the one you are used to injesting. It may even have its own agenda - depending on how the information is reprocessed. And it also has a shelf life since humanity continues to have new insights and people keep getting new alcohol brands to try.
“A government audit of family registries in Japan in 2010 uncovered more than 230,000 people listed as being aged 100 or older who were unaccounted for, some having died decades previously.
The miscounting was attributed to patchy record-keeping and suspicions that some families may have tried to hide the deaths of elderly relatives in order to claim their pensions“ Thus they need to redo the audit in 2026.
Looks useful, but the Mac download shows "“Capsule” is damaged and can’t be opened. You should eject the disk image." and then you realize it could be ripe for malicious payloads as well.
Preparing https://www.dreamlist.com for the holidays. It's a private wish list and registry site that enables you to also group lists of friends and family into one page to share with grandparents or social media. So you can run family holiday giving + gift drives / toy drives / disaster recovery with no ads or spam or any other badness.
When I first launched it was the only wish list that didn't list your name and registry on search engines. It is has grown a lot since then, and added a lot more privacy features, and the best marketing has always been doing right by users (extremely rare these days). You can see by looking up any name + the name of a wishlist/registry site and see what shows up to check at any time.
I'm working on more products following the same rule - serve users as you would your own family.
When a person spends time with other people, they outsource some of their thinking organically. That’s the efficiency gain in teams. Marriage leads to couples becoming two halves of a shared brain - one person takes on some cognitive load (finances, kids education, etc), while the other takes on another (maybe house issues, taxes, you name it). The load that can be carried with one brain is limited, so merging with others and outsourcing to others feels organic to the brain. LLM create the impression they have the cognitive ability to carry the load, but they don’t have the consistency of memory. You gladly assume you can finally get rid of a task and the model you trusted it to gets swapped under the curtain, or doesn’t train and improve on the task you outsourced, or the inference provider just quantizes it stupid. In the mean time you’ve lost your skill and your task has failed. Yes they feel good on some tasks some of the time, but still very much variably good (even same models different days of week or times of day). That’s closer to gambling and variable rewards than a spouse.
So if your code is spaghetti and that makes it impossible to add new features (happens easy, especially if you let an LLM loose on it), the refactoring to fix it would never happen and the engineers who stay or are willing to be hired will just be those clueless enough not to know they won’t be able to achieve anything.
Part 1: Everything was turned into social User Generated Content to reduce editorial costs. Part 2: Predatory third parties ranging from nation states to SEO agencies started mass filling UGC with anxiety inducing brain rot. Part 3: the brainrot generation is automated by AI and it silences any real people or voices. Part 4: People go outside to touch grass and Part 5: That leaves opportunities for platforms for real people with good culture
As with Cursor it's the data - all the prompts and responses that can be used to train AI models. Getting a head start on that at the pace of current development costs.
Let’s say a hacker hooks up the hotel internet to a firewalla and opens a free wifi spot. You log in and it sees from your traffic that you’re a customer of ABC bank and Gmail and ATT from the traffic - next thing you know you get a text message from your “bank” that someone has stolen your credentials in “XYZ Arizona” and you have yourself a phishing attack. The exposure scenarios are only limited by the imagination of humans and modern LLMs.
VPN is what a smart person uses when they are at an internet cafe or somewhere else with public internet where you can’t trust that the traffic isn’t sniffed by someone for giggles and your bank credentials are going to leak.
Completely agree, BUT you’re not dealing with -self-preserving humans anymore. In this case what are the odds of some less competent large state actor getting a model that goes rogue and has to be mitigated against? How do you mitigate against novel weapons it creates if your models are lobotomized in that category? No single government will neuter their own models and very few if any are competent and interested enough to mitigate against them.
A piece of wood can be used to build a house, to carve a statute, to make a bow and arrow, or to make a baseball bat. Honesty is the only way to see the problem in a way that can be tackled effectively. Framing will not help if you are dealing with large animals fighting in a cage when you are trapped in it, and it this case we all are.
The historic analog most befitting AI are weapons. It provides an advantage to companies and governments in creating weapons, in surveilling people and thus gaining leverage over individuals, in controlling what individuals can do or cannot stop doing. It thus follows the laws of weapons control - the more restrictions are threatened to be imposed, the more even the smallest actors will fight to get more of it. They may not understand or know what they are getting, but they will want more of it because their potential rivals have it. Governments are the ultimate customer here - they have the money, the energy, and the appetite to consume AI infinitely in battlement against opponents - foreign and domestic. The only approximate solution is to prevent any restrictions on the use of it - a-la-2nd amendment (the upside of that is it advances the AI for future use and is thus in the interest of governments as well). Modern large models seem to be compressible and thus we may see more and more tiny sub-specialized varieties. But the push to slow the frontier in weaponry always leads to the reverse.
Villages all over the world since ancient times would have houses with some form of a summer kitchen. During hot days the temps in the house are already uncomfortable and cooking there would make things worse, so they built basic kitchen facilities away from the main house to keep heat and smoke and smells out (that also reduced some fire risks).
Many western countries now summer-kitchen most of their uncomfortable output to corrupt poorer countries where the leadership doesn’t mind taking cash for poisoning their citizens. The naïveté is then claiming that those poorer countries are then the main polluters. Recyclables in the US and a lot of Europe are shipped off to southeast asia thus 80% of plastics in the ocean are found there. A train of plastic to be burned was even shipped to Bulgaria (a EU member) by the mafia a few years ago.
In the mean time “carbon credit” companies have their offices in Hong Kong and are guilting Disney and Adidas and Apple and any Western public brands into paying for “carbon laundering”. Forests that would have been planted anyway for lumber collect carbon credits as extra income right before selling said lumber for consumption. It takes real faith to keep believing any of the activity actually removes carbon legitimately, but the money movement in that movement is very real.
While anyone can target the source of the warming with any resource they want to dedicate, and should be free to do so. Governments need to focus on mitigating the risks to human life and that involves real fire prevention (trimming, controlled burns - maybe smart tech the smoke so it gets captured instead of irritating the noses of non-summer-kitchen citizens), clean energy generation, and installing damn air conditioning units into old European capitals so Europeans don’t end up dying in droves from preventable basic climate change consequences.
The dark ages and the plague saw a lot of belief in religion and rejection of science that lead to excessive deaths all over europe. Medicine was studied and practiced in the arab world and jewish doctors were the only ones allowed to bring some of those practices to Europe. Religion was preventing tested and verified approaches to treatment. Today Europe needs to stop repeating that practice. Believing in human caused climate change doesn’t stop you from learning proper tested mitigation strategies against wildfires, cooling of people in cities, nuclear energy or anything else that allows you to mitigate the effects of your climate religion. Because you can mitigate death in your area way before you can stop china or any other high emitter from emitting CO2 on the other side of the planet from you and contributing to global warming.
I’ve started evaluating what seem like brain “modes”. 1. Constant flow of information with disjointed context (social media, brain fracking, moving in a crowd) triggers hunter-gatherer instinct that is hard to stop. The dopamine seek in a way. Even if you stop your brain seeks it, so unless you really need this mode for work, it helps to postpone or time box these to the end of the day (kind of like some would do with drinking alcohol).
2. How you start your day after sleep creates framing of the brain and you can start iterating on frames and how you drive them depending on your needs. Maybe set a schedule of things to at need done the night before (a continuity thread) maybe boxing the activities to different types at different times. It’s all sciencing the self but the brain is definitely conditioning to things and you have to mind it.
3. ADHD meds from what i’ve read are caffeine on steroids (blocking adenosine receptors, but not really adding to executive function). From personal observation ADHD also feels like a mode one can be put into by wrong brain framing (at least for people not diagnosed).
The content within the models might be the play. If inserting the right content for the rest of the world to consume from the models is important to them, they will give away all the content they want the world to have.