It'd be fun to explore creating a Gemma 4 LLM API server app so you could use your iPhone's processing for agentic coding on a laptop. I don't know how useful it would be, but it'd be fun.
I want local models to succeed, but today the gap vs cloud models still seems continually too large. Even with a $2k GPU or a $4k MBP, the quality and speed tradeoff usually isn’t sensible.
Credit to Google for releasing Gemma 4, though. I’d love to see local models reach the point where a 32 GB machine can handle high quality agentic coding at a practical speed.
Fwiw, the real reason we don't have 100+ GB GPUs is because Nvidia likes to segment their markets. They could sell the consumer cards with 200gb gddr RAM on it, they just know that'd eat into their enterprise offering which is quiet literally all their profit margin (which I may add is gargantuan as of 2025)
If there is one thing that that agents/LLMs have highlighted, it is how much credit those test writers do deserve. Teams that were already following a TDD-style approach seem to be able to realize value from agents most easily because of their tests.
The tests are what enable: building a brand new JS runtime that works, rewriting a complex piece of code in a different language (e.g. Golang instead of TypeScript) that is more performant for that task, or even migrating off of an old stack (.NET WebForms) to something newer.
It is pretty incredible to me that in the pre-LLM/agent coding world, creating a new high-quality JS engine or browser seemed like it would likely never happen again. But now, any large company could build one if they wanted to. In a single digit number of months no less.
There's many JS implementations out there. Quality kind depends on what you need, and there's some engines more or less complete in which quirks are supported.
And for example, v8 doesn't make much sense in embedded contexts
There are definitely plenty of other JS engines, but they aren't always up to date on newer JS features. I'm pretty sure this is the 3rd JS engine to fully support the Temporal API (even JSC hasn't shipped it yet).
Some of the expanded scope that I’ve done almost for free is usually around UX polish and accessibility. I even completely redid the —help for a few CLI tools I have when I would never have invested over an hour on each before agents.
I agree that the efficiency and quality are very hard to measure. I’m extremely confident that when used well, agents are a huge gain for both though. When used poorly, it is just slop you can make really fast.
I'm a sucker for this kind of stuff, but I'm usually disappointed after clicking the link. These on the other hand are excellent, and that they have configurable options like stroke, color, etc is gravy on the top. Thanks for sharing!
This is exactly what happened at a SaaS company I previously worked at. It was an awesome company with ~1500 employees, turning a small profit. Private Equity comes along, buys it with ~$2B in debt. Sticks the SaaS company with a $100M+ annual interest payment. Round after round after round of layoffs ensued. Then interest rates went up... and it got even worse.
I think they are under 500 employees now. They basically laid off almost all of engineering and hired 100 new contractors in India to completely rebuild the entire platform in Node.js, as if the language it was written in was the problem. So glad to be far from that dumpster fire.
Really disappointing to see a great company gutted by some private equity people who almost certainly got their bonuses before the shit hit the fan.
I don't understand: Who's lending the $2B in situations like this? Wouldn't they be worried that the above situation (company gutted, then going down the drain) is going to play out and they won't get their $2B back? Or is that the root problem with this whole YC submission: banks are being hit by defaults because of this exact problem?
It's from the rapid exploitation of an asset. If I have a cow, I can milk the cow or kill the cow. If a cow costs $1, maybe I can get $5 worth of milk over the cow's lifespan, or I can kill the cow immediately and get $2 of meat. The man with $100 who buys all the cows in town and kills all of them doubles his money in a short timespan, but now there's a shortage of both meat and milk next season.
That is the exact problem. The people who put up the $2B thought they were safe - after all they were putting up money to buy a successful profitable business.
Problem is they didn’t really understand the business and trusted the PE guys to keep running it well…
This was driven home to me at SaaS company with > $80M ARR when the new CEO was parachuted in by the PE owner said in an all-hands "and we're close to cashflow positive when we account for our interest payments..." How can a software company generating this much subscription revenue NOT be making money? When it's servicing the > $500M the PE firm used to buy it. The rest of the playbook was boringly predictable: cut costs, sign multi-year enterprise deals, sell before the current fund's horizon and hope the music doesn't end.
As a result I prefer the naked greed of VCs where everybody - VC, owners, employees - knows the plan is IPO because at least it's transparent compared to the dirty lies a lot of PE pushes.
It's the destructiveness that gets me. It's a perfectly good company, employees are happy, consumers are happy, profit is being made, it's sustaining itself... Then they come and just literally destroy all that.
This can't be good for society. I wonder why it's just not criminalized somehow.
> It's a perfectly good company [...] I wonder why it's just not criminalized somehow.
Not-an-expert here, but I think part of the problem is that it's hard to draw a nice legally-enforceable line that would distinguish when it's a "perfectly good" company versus one crying out for intervention.
For example, suppose a company is floundering because of executive mismanagement, outrageous compensation to the C-suite, etc. In that case, someone could LBO in, fix things up, and then sell the revitalized thing later and make a modest profit while improving the world.
The way I see it, it's literally simply the PE paying the existing owner for the privilege of squeezing the value out of the business and its customers in the short term (or in the ideal/theoretical case, running it more sustainably and making higher profits). Management's job becomes to extract high profit in the short term, not to keep the company running profitably.
So, logically, selling to PEs/operators who are known to do this is basically the owners selling out and taking the cash. The consequences are clear to anyone who's been watching.
I saw another model where the PE buys a hospital. They sell the land under the hospital, everyone gets a cut, then they spin out the hospital. Now the hospital has to pay rent on the land it sits on.
It seems like almost every decision made is for short term gain, at the cost of long term viability.
It's funny that "restructure the codebase to be more friendly to agents" aligns really well with what we have "supposed" to have been doing already, but many teams slack on: quality tests that are easy to run, and great documentation. Context and verifiability.
The easier your codebase is to hack on for a human, the easier it is for an LLM generally.
Turns out the single point of failure irreplaceable type of employees who intentionally obfuscated the projects code for the last 10+ years were ahead of their time.
I had this epiphany a few weeks ago, I'm glad to see others agreeing. Eventually most models will handle large enough context windows where this will sadly not matter as much, but it would be nice for the industry to still do everything to make better looking code that humans can see and appreciate.
It’s really interesting. It suggests that intelligence is intelligence, and the electronic kind also needs the same kinds of organization that humans do to quickly make sense of code and modify it without breaking something else.
reply