That's almost always been true but servers are more like commodities at this point.
iphone is nice and upgrade from commodities phone and I have one, but I wouldn't care much if my fridge has sleek UI because I just need it to be a fridge.
I was neovim in the end, 100% agree lua is so much better than vimscript, but now I don't need either. I spend no time trying to match what an IDE can do in the terminal and get to spend that time building the things I'm actually interested in. I recalled Linus saying the reason he (at the time) used Fedora was because it just worked and he could spend his time on the kernel instead of tinkering to get linux working. This is one of the biggest reasons I stopped using (neo)vim
I had lots of problems with plugins in the ecosystem breaking, becoming incompatible with others, or often falling into unmaintained status. Integrations with external SaaS services are much better too
Also information density (and ease of access) as a peer comment has mentioned
> match what an IDE can do in the terminal and get to spend that time building the things
This is a common complaint but I haven't done any setup for months.. And installing a language server because I need to write typescript is just <leader>cm and then lllll on the servers I need.
My experience is the opposite. Terminal UIs make better use of information density because they don’t have ridiculous padding between widgets and other graphical chromes that modern GUIs have.
Also terminals support mice. Have done for literally decades.
Ultimately though, it just boils down to personal preference
Modern terminal emulators run at native resolution and support window splitting. You can have the exact same information density (I'd argue that a nice neovim environment has more information density than most IDEs since vs code and jetbrains seem to love putting extra space and padding everywhere now.)
Terminal emulators can only do exactly that. Emulate increasingly large terminals. They are almost fundamentally unable to render something smaller than a single character.
The big question - which one of these new agents can consume local models to a reasonable degree? I would like to ditch the dependency on external APIs - willing to trade some performance in lieu.
I spent at least an hour trying to get OpenCode to use a local model and then found a graveyard of PRs begging for Ollama support or even the ability to simply add an OpenAI endpoint in the GUI. I guess the maintainers simply don't care. Tried adding it to the backend config and it kept overwriting/deleting my config. Got frustrated and deleted it. Sorry but not sorry, I shouldn't need another cloud subscription to use your app.
Claude code you can sort of get to work with a bunch of hacks, but it involves setting up a proxy and also isn't supported natively and the tool calling is somewhat messed up.
Warp seemed promising, until I found out the founders would rather alienate their core demographic despite ~900 votes on the GH issue to allow local models https://github.com/warpdotdev/Warp/issues/4339. So I deleted their crappy app, even Cursor provides some basic support for an OpenAI endpoint.
> I spent at least an hour trying to get OpenCode to use a local model and then found a graveyard of PRs begging for Ollama support
Almost from day one of the project, I've been able to use local models. Llama.cpp worked out of the box with zero issues, same with vllm and sglang. The only tweak I had to make initially was manually changing the system prompt in my fork, but now you can do that via their custom modes features.
The ollama support issues are specific to that implementation.
LM Studio is probably better in this regard. I was able to get LM studio to work with Cursor, a product known to specifically avoid giving support to local models. The only requirement is if it uses servers as a middle-man, which is what Cursor does, you need to port forward.
it will add the feature, I saw openAI make that claim that developers are adding their own features, saw Anthrophic make the same claim, and Aider's paul often says Aider wrote most of the code. I started building my own coding CLI for the fun of it, and then I thought, why not have it start developing features, and it does too. It's as good as the model. For ish and giggles, I just downloaded crush, pointed it to a local qwen3-30b-a3b which is a very small model and had it load the code, refactor itself and point bugs. I have never used LSP, and just wanted to see how it performs compared to treesitter.
The discussion here reminded me of a funny (but probably true) anecdote where all you had to do to actually get working answer from stackoverflow (RIP) is to create an answer that is blatantly wrong.
You'll have people of knowledge descending on that question in no time to refute your heresy.
A proper ebike won't stand a chance against the modern queen stage of the tour de france, even if ridden by a professional with appropriate gears otherwise, because the battery would run out half way on the first HC and it would just be a very heavy bike for the rest of the stage.
Same with a tiny motor - you gain tiny amount of force but you'll have to carry a full bidon with you on all the climbs, not to mention that the delicate mechanism can break easily.
The bikes have a weight regulation that was set in the 90s, 6.8kg.
Ultra light bikes can be as light weight as 2.7kg. That gives 4kg to hide a battery and motor and still hit weight. A really good lithium battery offers 350 Wh/kg. 1kWh can grant 100 miles of range by itself.
That weight comes obvious in components. All teams are required to use widely available components and it's quite easy to spot one that's not normal. For the bike builds that are 4kg or even less, it's quite obvious that all components are non-standard.
You can save at max a bidon before rousing suspicion, and the whole operation is just not feasible in terms of cost vs. benefit.
> the whole operation is just not feasible in terms of cost vs. benefit.
Batteries and a motor are a huge benefit. Even if you can't squeeze in a full blown motor or 1kwh of battery, just getting an additional 200 or 300 kwh of assist in can make a huge difference.
As for cost, these guys are already doing crazy things like blood doping just to get a tiny edge.
By contrast, blood doping is much easier to get away with. Claim you had an altitude camp, inhaled CO or whatever, but you carry your blood with you.
A small motor had to fit in the tubes, somehow connect to a control, have to be integrated into the gearing which are constantly under about 300 W of torque and can be easily discovered via X-ray or maybe heat gun. That's a lot more risk vs. a much smaller reward since your laptop sized battery is likely less juice than a single energy gel.
While I don't believe they're being used to cheat in professional cycling, a motor would _definitely_ provide a massive advantage in a cycling race of any kind.
A motor easily provides enough power to overcome its weight, and they wouldn't need assistance for the entire race, just an edge at key moments.
Think of the riders themselves as incredibly efficient batteries and motors - they can also recharge at 120g carb/hour. The motor itself is just deadweight over most of this process.
But the weight doesn't matter most of the time - on flat sections and downhill, which are 90% of the distance covered, it's completely irrelevant.
For much of the stages, the top guys are not doing much work, they spare their legs for the climbs. They will hide in the pack, doing only very light work drafting. If you could put a smallish battery able to recharge on flat / downhill sections and only provides a boost on the critical uphill parts, that would be a massive advantage.
A lot hinges on this magical contraption that can both accept connection and command, deliver power in sufficient amount when needed while being integrated into the hub, light enough to hide in a bike and somehow even be able to recharge. Have you seen the ebikes that they sell and the hubs that they have?
Now it could definitely work if you only care about a single attack - not even a full climb, but maybe 200-300m where you want to distance your rivals, but the risk (not just of discovery, but of malfunction, etc) just doesn't seem to be worth it. There's domestiques for a reason.
Its real and I find it technologically fascinating as they were using the frame and wheel as motor.
In January 2016 – almost six years after initial allegations of a pro cyclist
doping mechanically – the first confirmed use of "mechanical doping" in the
sport was discovered at the 2016 UCI Cyclo-cross World Championships when one
of the bikes of Belgian cyclist Femke Van den Driessche was found to have a
secret motor inside. One blogger described it as the worst scandal in cycling
since the doping scandal that engulfed Lance Armstrong in 2012.
I think that's just a natural progression from the fact that doping was the main way to cheat in the past, so "mechanical" doping is just the new doping.
But also: no one's ever thought doping wasn't cheating anyway. It's certainly not a euphemism in cycling.
You wouldn't necessarily use mechanical doping to win the general classification, or even a particular stage.
More likely, you'd use it on select stages for very specific reasons... for example, a rider could use it to avoid the time cut on an ITT stage (effectively getting extra rest vs their competitors). Similarly, a pure sprinter could use it to stay in contention on a punchy "sprint" stage (like a stage that MvdP might be a favorite instead of a pure sprinter).
Edit - I don't think anybody is doing this at the top levels of pro cycling. Maybe in regional racing (masters, etc).
If I were responsible for a mechanical doping program, then I'd install the motors for the leadout and mountain domestique riders and leave the team leader clean. Who cares if they pay the weight penalty after peeling off if it means that they can provide extra support for those critical minutes?
> would just be a very heavy bike for the rest of the stage
Bikes in the Tour de France have a minimum weight of 6.8kg imposed by the UCI. So if you manage to build a normal bike that weights 5kg, you still have 1.8kg of weight available to try to add some more hidden power "without adding more weight to the bike" (small battery+engine, small compressed air tank, whatever).
Funnily enough, you're correct in your belief, even if by accident and in defiance of your own preconception. Mechanical doping is the topic your speaking about! :)
Here's some of the more obvious examples out there:
A hybrid car trivially improves total energy input needed, since it replaces braking by generating heat by braking by storing energy later to be reused.
The same should he true here, right? The added energy needed to carry the weight of the motor would be easily overcome by the gains from regenerative braking?
Only if the motor were in the hub of the wheel, which given the typical size of the hubs, seems even less likely. Remember that bicycle drivetrains are typically one-way due to the ratchet, so you can't apply braking force via the chain.
Are you saying the physics of a bicycle are somehow different than a car going up and down hills? Or are you saying actually hybrid cars use more gasoline driving in hilly terrain as well, and their benefits only accrue in stop-go city traffic?
Physics and practical concerns are way, way different. You want to go as fast as possible down the descents in a bike race. You don't want to lose any kinetic energy and fall behind your opponents, so the only time you'd be using it is when you actually want to slow down. In a car, you might be braking/slowing down going downhill anyway, so that energy is better captured than used that moment.
There's also the matter of mass: lot more momentum/energy to be gained from a 1500kg car versus a 70kg bike + rider. That said, less energy needed for the motor so don't know how the math works out there.
Edit: all of this is moot anyway because of the point zettabomb made as well.
It would be extremely beneficial, but nearly impossible to integrate. Motors used for cheating in cycling are usually in the seat tube or down tube, where they can invisibly interface with the bottom bracket (between the pedals) and connect to batteries elsewhere in the frame. Because bicycles have a freewheel in the rear hub (chain doesn't move while coasting/braking)*, a KERS would have to be located in the tiny rear wheel hub.
*You can of course get a non-race bike with a fixed chain, but UCI rules require use of a freewheel.
There is no way to do this unless the motor is inside the wheel hub and that would be instantly obvious - regular hubs are super thin and wouldn't fit a motor + capacitor inside them. And you'd need to tell it you want to brake somehow.
but cycling races are won by being able to put out a critical extra 50 watts for a few minutes at a key point in the race. I don't think anyone is trying to motor the whole way up a climb, but I can imagine how you could have a useful motor if you're just trying to run for ten minutes total? at that point it's analagous to the <250g drones that are out there.
I'm loving it, can't wait to deploy this stuff locally. The mainframe will be replaced by commodity hardware, OpenAI will stare down the path of IBM unless they reinvent themselves.
iphone is nice and upgrade from commodities phone and I have one, but I wouldn't care much if my fridge has sleek UI because I just need it to be a fridge.
reply