Yes, it's true, if you continually invested and averaged down over the last 2years your money has slowly deteriorated. While I don't believe in the company reaching any absurd total market cap, like $5,000+ per share, I do believe she's got the chance to run up to $100-500 per share, at which point those 3-4 years of waiting for something to happen was well worth it. The company has $1b and no debt, has turned profitable, has $100m on deck to buy back stock... Sure, it's no tech giant, but it's still extremely undervalued. If it was undervalued at $4 pre-split when DFV was talking about it and the company they were at the time, it sure as shit is worth at least $40 pre-split with how the company is doing today.

---

How fuckin poor are you? 60k pre-tax for a family is like fucking destitute. Like idk how you live off of it unless your entire life revolves around a budget and garbage picking. Blows me away how poor people are especially on reddit, I guess that's why they have antiwork subs and everyone wants free money.

---

> It failed back in 2008 and the company should have been dissolved What are you talking about?!? GM WAS dissolved in 2008!! The GM of pre 2008 no longer exists, that company went bankrupt & stockholders lost their shirts. I have family members who were very impacted by this reality. Then the company was reformed with the remaining assets as a government owned entity, and shortly after refloated on the stock market. TOTALLY destroying General Motors would not have been a good idea. The model used for the GM bankruptcy is, what I would argue, exactly what the government should do w/companies that are 'too big to fail' but that need to be destroyed. This is exactly what should have happened to all the banks that caused the 2008 crash, instead of handing them money to give out as bonuses to the 'talent' that where really criminally liable.

---

This is some old school, pre-GME wallstreetbets autism. Love it.

---

Bro WSB was exactly that type of sub for years beforehand. Sure, there were still lots of guhs and money in gourd futures moments between 2017-2018 when I started lurking here too. However, for years and years, wsb was focused on talking about actual positions and money making while laughing at the people who fucked it up. Wsb right now is actually closer to the pre 21 subreddit than at any other time I can think of after Gamestop. That shits never going back but at least it's bearable to actually browse the sub again.

---

I was on WSB pre sneeze. I haven't been here for the past few years. Why can't we mention Gamestop here? Seriously asking.

---

It's about both, amount of data (and quality as well) is very important with pre-training, quality is the main thing with alignment/fine tuning. That's my understanding, at least. So at some stage, you need that initial data to train the model, or to train the model which generates your synthetic data. And you need a lot of it. Also synthetic data can be very useful, but for obvious reasons you can't really start there, unless you do what everyone does and just use gpt-4 to generate data for you, but openai isn't too happy with that and will probably notice if you make billions of api calls generating synthetic training data for your competing model. This applies mostly if you're the one training the base model, so if you're openai or meta. If you're just doing a fine tune of LLaMA, as many of the AI companies do, you just have to care about the fine tuning data, and will have an easier time generating synthetic data, since you need a lot less of it. And I would guess LLaMA-2 might be good enough to make a ton of synthetic data for many use cases as well. I would think that the licensing of that model allows for this, but I'm not sure.

---

Tulips kinda pre-date him by a few years, and don't get me started on gourds...

---

Yes a pre market pump so surprising. When does this even stop, clownery . At this rate its going to be 1% by open. Regarded

---

Just gone green in the pre-market. Hope we're not both regarded OP.

---