Justin Beck

Just a simple Quant

[Justin Beck @justinbeckcoin]

@LottoLabs no context, is it free or we buying it? Free 100% get the 6000, sell it, buy 4 sparks.

[Justin Beck @justinbeckcoin]

@MiaAI_lab poor husband getting downgraded LOL

[Justin Beck @justinbeckcoin]

@Xavier_Sanchez4 This randomly popped up in my feed and I'm just over here like... LOL people actually argue about WNBA.

[Justin Beck @justinbeckcoin]

@Umesh__digital ignorance.

[Justin Beck @justinbeckcoin]

@EvanLuthra have to wait 120 hours for my reset rip

[Justin Beck @justinbeckcoin]

@Abomination81 As Nancy Reagan said in the 80's. Just say no. Drugs, money, same difference.

[Justin Beck @justinbeckcoin]

@ZssBecker @UziCryptoo Don't assume other people feel like you. We are all different. You're judgements and opinions of what other people feel and want is nothing to us but more internet garbage. Spend less time thinking about other people and more with your kids if that's what you care about.

[Justin Beck @justinbeckcoin]

@perplexity_ai Handing over all of my personal info including ssn to AI seems like such an amazing idea.

[Justin Beck @justinbeckcoin]

@ZssBecker Tell us you're bad at vibe coding without telling us you're bad.

[Justin Beck @justinbeckcoin]

@michaelrbock @Hail_BAIJ @_sportscardshop @claudeai @turbotax @AnthropicAI don't take marketing advice from this guy folks.

[Justin Beck @justinbeckcoin]

@Abombination81 @dayal631 @RblxSukuna ngl had to laugh a little at that.

[Justin Beck @justinbeckcoin]

@LakeBoss ok guys, what game was she playing?

[Justin Beck @justinbeckcoin]

@bygregorr @IndieGameJoe While i don't know the specifics of why he chose to go with making his own engine, he did say some of the mistakes he made was going with his own engine and spending a lot of time with vulkan and openGL before using DirectX. Those things made the game take 4 years to make.

[Justin Beck @justinbeckcoin]

@kartikktwt until you figure out quickshell doesn't support many things a developer needs to have. It's missing so much that it's unusable for anything other than vibing.

[Justin Beck @justinbeckcoin]

@austingriffith ram means very little in the AI world. You need GPU's that can handle the workload with as much VRAM as you can get. Without that it's going to be slow and painful.

[Justin Beck @justinbeckcoin]

@Tnimbus @MoonDevOnYT Nope, it's impossible. These would struggle even with the heavily quantized gguf versions of kimi.

[Justin Beck @justinbeckcoin]

@devmaxie_ @MoonDevOnYT It's not openclaw that needs the power, it's running local models with a GPU. There are better ways to do this than a mac pro, but it's easier for the average person.

[Justin Beck @justinbeckcoin]

@MoonDevOnYT Depends what you're doing. OpenClaw can do more and has more use cases.

[Justin Beck @justinbeckcoin]

@hellonatch Plus, I didn't have to learn new skills to set this up. Sometimes buying mac minis or a mac pro is worth it just because of the work it would take for someone to learn and set up a system like this.

[Justin Beck @justinbeckcoin]

Mac minis for openclaw are not the way to go if your are serious about this tech. I can literally spin up an unlimited amount of ubuntu VMs on my home server way cheaper than you can purchase your 3 mac minis. GG.

[Justin Beck @justinbeckcoin]

@hellonatch I still use APIs for the brain of my operation, but my hypervisors have GPU passthrough setup allowing me to run large LLMs locally.

I luckily had the setup before the RAM epidemic.

Mac minis are a decent option to get you going fast with the current price of hardware.