Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
SwellJoe
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
24 ms
·
571.
▲
by
SwellJoe
5mo ago
No. The insiders don't need it to go up from the IPO price. It can fall quite a bit, and they still win. Your ETF will likely be catching a falling knife. All the funds that track the Nasdaq 100 will prop up the price so it doesn'
572.
▲
by
SwellJoe
5mo ago
SpaceX is an even bigger scam, if that's possible.
573.
▲
by
SwellJoe
5mo ago
Not every company trades at 387 PE on 2%-3% growth. I can't name any that do. F sells an order of magnitude more cars (~$190 billion in revenue), and has a market cap of $62.8 billion vs TSLA market cap of $1.59 trillion. TSLA is ridic
574.
▲
by
SwellJoe
5mo ago
It is reasonable to predict TSLA stock will, someday, drop to something roughly reflective of the value of the company. It has a long way to fall to reach that point, but, there's no reason to think it won't happen eventually. The
575.
▲
by
SwellJoe
5mo ago
I don't care if it's fast, if it makes the model dumber by cluttering up context.
576.
▲
by
SwellJoe
5mo ago
Qwen 3.6 35B-A3B with MTP at 8 bits is blazing fast, something like 50-60 tokens per second. That's plenty fast for interactive use, so I haven't tried lower bits. Unfortunately the MoE is notably dumber than the dense model (for
577.
▲
by
SwellJoe
5mo ago
The hate is far less than is warranted. This is purely a transparent transfer of wealth from working people to the richest people on earth. Elon Musk demanded the right to dip into average folks retirement funds for himself and his investor
578.
▲
by
SwellJoe
5mo ago
Even batched it's uncomfortably slow. I started to benchmark ds4 with my security vulnerability benchmark (after Qwen 3.6 dense and MoE and a bunch of cloud models), but it was going to tie up the Strix Halo for more than a day, so I d
579.
▲
by
SwellJoe
5mo ago
I haven't, but, probably will add it to the benchmark soon.
580.
▲
by
SwellJoe
5mo ago
After working with LLMs a bunch, I now want them to forget everything every time I end the conversation. Otherwise they get dumber and more confused over time. LLMs do not have memory and these "memory" systems that everyone mak
581.
▲
by
SwellJoe
5mo ago
Yep, I have a Strix Halo and while it can run models bigger than Qwen 3.6 27b, it's not usable interactively when you do. ds4 patched for ROCm works, but at such a slow speed, it's not usable for coding agents. The Nvidia boxes
582.
▲
by
SwellJoe
5mo ago
> So one of thr things you're using it for is to generate leads to spam businesses with unwanted LLM produced marketing materials it sounds like. You don't know me. And, no.
583.
▲
by
SwellJoe
5mo ago
Everybody builds one. And, then they usually figure out that making the model fill its context with a bunch of memories hurts performance more often than it helps.
584.
▲
by
SwellJoe
5mo ago
They already provide E2B and E4B that run on (much) smaller devices, including tablets and phones. This fills the gap in the middle. The bigger Gemma 4 models are excellent for their size, but at 8-bit quantization they need about 64GB of V
585.
▲
by
SwellJoe
5mo ago
I mean, they did invent the technology. It's actually kinda surprising they're not the leader in the space. They kinda got Kodaked (though the story is still playing out, and I guess they're still somewhat competitive in the
586.
▲
by
SwellJoe
5mo ago
I've used Gemma for reviewing and categorizing my writing online over several years (~5 million words across a forum for an OSS project I work on, HN, reddit, etc.), experimenting with training LoRAs (again, on my own writing, since I
587.
▲
by
SwellJoe
5mo ago
At $3/$15, Sonnet is more than an order of magnitude more expensive than DeepSeek at $0.435/$0.87 (with cached input pricing of $0.003625, DeepSeek is very good at caching, so it's very cheap to use). So, if they're equa
588.
▲
by
SwellJoe
5mo ago
Sure, but it's priced higher than many better models. I'm not saying use the biggest models for everything. I'm saying Haiku is not a great deal as small models go. You can even self-host a model that is competitive if you&#x
589.
▲
by
SwellJoe
5mo ago
I've been doing benchmarking of various models for finding hard security bugs, and my faith in Haiku (and Sonnet, even) has dropped precipitously in the process. Self-hosted Qwen 3.6 27B consistently outperforms both for finding securi
590.
▲
by
SwellJoe
5mo ago
I assume not all indexes will make this deal with the devil, as Nasdaq is doing. At least, not immediately. But, given how lax regulation has become and how corruption and lies from the biggest companies has become commonplace, I have to as
591.
▲
by
SwellJoe
5mo ago
It is a liquidity event for Elon Musk, and people like Elon Musk. What they do with it, hard to say. But, the forced buying by institutional investors will push the price of SpaceX upward based on nothing. SpaceX revenue was $15.8 billion l
592.
▲
by
SwellJoe
5mo ago
Bogleheads aren't wrong, historically. At least, not in the general sense that buying index funds and mostly forgetting about it is smarter than trying to beat the market with individual stock picks and timing the market. But, that phi
593.
▲
by
SwellJoe
5mo ago
I'm informed enough to see what they're doing and why, but I'm not informed enough to know how to prevent them from wrecking the economy for normal folks while enriching themselves. I don't think information alone can so
594.
▲
by
SwellJoe
5mo ago
Which, again, benefits the wealthy and well-informed and well-connected. The suckers who have their retirement savings in some kind of index fund because all the experts have been saying, "Buy index ETFs and forget about it" for d
595.
▲
by
SwellJoe
5mo ago
Because it's a scam by the richest people in the world to steal from the retirement accounts of everyone else.
596.
▲
by
SwellJoe
5mo ago
Housing values did not tank in Fort Worth. You can't just make up the results to suit the story you're telling. Fort Worth housing prices have mostly tracked alongside the rest of the country (big spike upward starting around 2020
597.
▲
by
SwellJoe
5mo ago
I think the value of it is currently more academic than useful in the real world. Everything at the frontier is still only marginally Good Enough (in image generation, most of it is shit even from the best models), so things far behind the
598.
▲
by
SwellJoe
5mo ago
There's the normal level of... optimism. And, then there's SpaceX, a scam on a scale the world has never seen.
599.
▲
by
SwellJoe
5mo ago
With the new Nasdaq 100 fast track rule, I'd certainly get out of any index that tracks it, or any funds that are invested in it. I don't know if any other indexes have had similar policy changes...but, if it works this time, and
600.
▲
by
SwellJoe
5mo ago
The company is wildly overvalued. It'd be funny if Musk wasn't about to walk away with a bunch of money stolen from retirement accounts.
More ›