Well it raises an interesting conundrum. Suppose there's a microcontroller that's $5.00 and another that's $0.50. The latter is a clone of the former. Are you better off worrying only about your short term needs, or should you take the long view and direct your business towards the former despite it being more expensive?
Suppose both microcontrollers will be out of date in a week and replaced by far more capable microcontrollers.
The long view is to see the microcontroller as a commodity piece of hardware that is rapidly changing. Now is not the time to go all in on betamax and take 10 years leases on physical blockbuster stores when streaming is 2 weeks away.
Ai is possibly the most open technological advance I have experienced - there is no excuse, this time, for skilled operators to be stuck for decades with AWS or some other propriety blend of vendor lock-in.
This isn't betamax vs VHS. It's VHS vs a clone of VHS. The long view necessarily has to account for R&D, long term business partners, and lots of other externalities. The fact that both manufacturers will have a new model of VCR out next month, and yet another the month after that, really has nothing to do with the conundrum I gave voice to.
I'll also note that there's zero vendor lock-in in either scenario. It's a simple question about the tradeoffs of indirect parasitism within the market. I'm not even taking a side on it. I don't even know for certain that any given open weights Chinese model was trained against US frontier models. Some people on HN have made accusations but I haven't seen anything particularly credible to back it up.
If the clone is 1/10th of the price, and of equivalent quality, why would I use the original ? I would be undercut by my concurrents if i did that, it would be a very bad business decision.
Well the company of the former microcontroller has gone out of their way to make getting and developing on actual hardware as difficult and expensive as possible as possible, and could reasonably accused of doing “suspect financial shenanigans”, and the other company will happily sell me the microcontroller for a reasonable price. And sure, thy started off cloning the former, but their own stuff is getting really quite good these days.
So really, the argument pretty well makes itself in favour of the $0.5 micro controller.
That's a very tenuous analogy. Microcontrollers are circuits that are designed. LLMs are circuits that learned using vast amounts of data scraped from the internet, and pirated e-books[1][2][3].
Hmm, that benchmark seems a little flawed (as pointed out in the paper). Seems like it may give easier problems for "low-resource" languages such as Elixir and Racket and so forth since their difficulty filter couldn't solve harder problems in the first place. FTA:
> Section 3.3:
> Besides, since we use the moderately capable DeepSeek-Coder-V2-Lite to filter simple problems, the Pass@1 scores of top models on popular languages are relatively low. However, these models perform significantly better on low-resource languages. This indicates that the performance gap between models of different sizes is more pronounced on low-resource languages, likely because DeepSeek-Coder-V2-Lite struggles to filter out simple problems in these scenarios due to its limited capability in handling low-resource languages.
It's also now a little bit old, as with every AI paper the second they are published, so I'd be curious to see a newer version.
But, I would agree in general that Elixir makes a lot of sense for agent-driven development. Hot code reloading and "let it crash" are useful traits in that regard, I think
Does anybody know how much an ML model is actually worth to build a new model? Like when they start making a new model, do they modify the old or do they start from scratch?
I'm asking to know how much owning a model is actually worth, not in how much it could make money by selling use, but in how much it deprecates and keeps value to make a new one. If say one side of China/US lacks out on a model generation, do they only need to follow progress on the science behind it and when they own the data, the algorithm and the hardware all they need is "just" time and energy or is it important, that they actually have their on instance of a large model from every generation continuously?
In high school I knew a kid who would go around looting loose change from unlocked cars. He'd pull the driver side door open like it was his car, hop in, loot the center console, then hop out like nothing happened. He wouldn't take valuables (as far as I knew), just change, so maybe a few bucks per car.
His rationale? "Nobody will cry over a few missing quarters and they are free to lock their doors anyway."
The reason it's not stealing is because the cost to the serve content is tiny (spare change) and the sites don't stop you from viewing it with ad-blocker (unlocked doors).
