Let me clarify: 10-20x is not going to happen "soon" because there are several competitors who are more than willing to take business.
They are already raising prices. Currently, paid demand is outstripping capacity.
For people who work in-town (a lot of trade contractors) they often don't even need charging infrastructure if they can charge at home. 200-300 miles of city driving is enough for a day of driving & night of charging.
Absolutely true, and the number of 1-ton trucks is excessive.
However, I stand by my take the Lightning was smart "meet them where they are" and good cost cutting to reuse an existing vehicle for Ford's first attempt.
They quickly started developing a new, smaller EV platform so they're doing both.
Totally agree about city errands.
Contractors/builders need work trucks.
Plus I wish there were more EV van options than the E-Transit & Rivian EDV.
The reuse of the platform was very smart from a cost perspective and "it's the truck you know and love but no gas".
Looking forward to Ford's next EV truck later this/next year.
Mechanics aren't stupid but also don't know everything. They can learn and also have subscriptions to and regularly use specialized repair manuals for every vehicle.
There's not much to learn for an EV. They've been doing hybrids for 20 years and there is little difference for an EV.
What do you mean by "cost more to repair"?
- Battery repairs are $10-$20K but lifespan of a battery is ~20 yr. And retain massive core value
- regular maintenance is massively reduced
- ICE complexity is increasing which reduces reliability and increases repair expense
Plus the big charger site brands ( @ionnaofficial.bsky.social @electrifyamerica.com ) claimed NEVI funding was a bonus but never relied on that to make a site profitable. NEVI had restrictions so it would never work for all sites.
I'd be interested in an honest take on Ford and GM's low key continuation of EV platforms. GM's platform-formerly-known-as-Ultium and Ford UEV are continuing. Cadillac is 30% EV. The big companies are a mix of short and long term thinking.
Plus Ionna, Walmart are accelerating charger installation.
People often call me Taylor but that swap makes a lot more sense.
We as humans can spot these inauthentic comparison articles, so it's possible the LLMs will get good enough to spot them too.
Perhaps I'm too optimistic.
Isn't this also what Open Code was doing a few weeks ago? I thought it was banned then.
Any examples or patterns?
Needs to be seen on the site, has a wonderful drop-down menu.
Just realized @doctorow.pluralistic.net is active on Bluesky and has an AI book coming out in 3 months.
Plus his long-time anti-copyright stance.
Even Doctorow has a nuanced and reasoned take, he uses Ollama locally: pluralistic.net/2026/02/19/n...
Which makes sense as I believe he is a technologist who is anti-corporate.
Haha no, I truly appreciate the perspective.
This is ok for an attacker! Spin up multiple instances, let them churn for hours, days, weeks. Like a thousand monkeys on a thousand typewriters.
Absolutely. Good practices are now more important than before.
It doesn't necessarily make the code better, but the same tools that malware writers can use are now easier than before when red-teaming for defense.
I'm concerned about this: AI gen can write code against the API where no malware writer would have spent the effort to do so before. It automates a lot of processes that were not worth the time before.
Yeah, I agree. The chance of a vulnerability goes up the more surface area/API exists.
The chance of any one of those being used goes up the more time it's been public.
AI accelerates the process of finding these but is not necessary.
ohhh yeah I'm re-reading what I wrote and I really kinda misstated what I was trying to communicate.
Consider malware: if you have a public facing service, it only takes one bad actor using AI to put your customers at risk. Not using AI to compete and protect against mal uses of AI will be extremely difficult and possibly malpractice. AI may end up being required for software security.
I'm open to being wrong!
I'm not here to be stubborn or closed off to learning.
Shitty websites predate LLMs and it's a business choice to spend more time making it right or to deploy something not fully tested.
I agree LLMs increase the pace of development and business expectations and the software industry is still adapting.
Are you familiar with software engineering and testing techniques such as red teaming, penetration testing, fuzzing, etc?
Attackers use these same techniques and LLMs can automate and do it faster than people. It is irresponsible not to think about, model, and protect one's products from abuse.
