Hacker Newsnew | past | comments | ask | show | jobs | submit | storus's commentslogin

Well, NAD+ kinda has the potential to fix multiple things at once if those things are caused by energetic imbalance/deficit. Re-balancing NAD+ could just fix multiple failing systems that were running low on energy and not doing their job properly as a consequence.

It's quite terrible how the medical "debugging" works or rather doesn't. You run a bunch (at best) of tests then pick the most probable diagnosis and that's about it for 99.9% of cases. And then in a review you measure that the world's best performing doctor hit 45% accuracy whereas an average one hits ~33%.

As someone who has been debugging their own chronic illnesses for the past ten years (and doing better than my doctors), I wouldn’t consider the medical diagnostic process to be “debugging” at all. And that is exactly the problem. Doctors seem to be stuck thinking like bureaucrats following probabilistic flowcharts, and they’re incapable of actually thinking about a problem and debugging it.

The behavior seems to be so deeply ingrained in every single doctor I’ve seen that it seems impossible to change. I suspect they must have this drilled into them in school and residency, then it seems like every decision is constrained by insurance requirements. As far as I can tell, the situation is hopeless.


NMN and NR are both good but NMN might not be available anymore as some company decided to repurpose it as a drug instead of supplement. Best combo nowadays looks to be liposomal NR with pterostilbene, a sirtuin activator. NR boosts NAD+ (the main electron transporter in mitochondria), pterostilbene activates sirtuin SIRT1 that regulates aging. B2/Riboflavin might be a good idea as well as it is a FADH donor, secondary electron transport carrier especially in nerves/brain. B1 to the mix as every single metabolic reaction needs it and it's depleted by consuming lots of carbs or drinking alcohol, a common western diet. Niacin is less effective in raising NAD+ but the flush can open up veins and flood hard to reach extremities of the body with blood so it's probably good from time to time. Slow B3 seems to be even worse for raising NAD+.

> the flush can open up veins and flood hard to reach extremities of the body

What does it mean if I don’t flush? Is the supplement a dud?


I know of three possible reasons:

- your supplement is a dud

- you took niacin too often (best to do it once a few days as body adapts quickly)

- you have a gene mutation that prevents you from absorbing enough B3 (common in some schizophrenia cases that can be managed by huge doses of daily B3, like 4-10g)


> you took niacin too often

Oh yeah, I was taking it daily when I stopped flushing. Makes sense.


Maybe he had some idea that his buddy ran away with and he attributed his success to his idea, while he was homeless? There was a whole TV show around a similar idea (Breaking Bad).

Im pretty certain that of plagiate software would include rejected students works whole swaths of the academic successful would be let go.

What were you actually trying to type? I can't figure it out.

I think the first "of" -> "if" makes it readable.

Why doesn't Cursor allow selecting a LLM for code completion in the UI anymore and forces "auto" everywhere now? I have a Pro account and noticed this started like a month ago, and the "auto" output was often garbage, not following the instructions.

I don't experience this. I'm still able to choose models with meta+/

M3 Ultra has a crappy GPU, somewhere around 3060Ti-3070. Its only benefit is the memory throughput that makes LLM token generation fast, at around 3080 level. But token prefill that determines time-to-first-token is extremely slow, and coincidentally all those tasks you mentioned above would be around 3060Ti level. That's why Exo coupled DGX Spark (5090 performance for FP4) with MacStudio and sped it up 4x. M5 Ultra is supposed to be as fast as DGX Spark at FP4 due to new neural cores.

I think the op meant pipeline parallelism where during inference you only transfer the activation between layers where you cut the model in two, which shouldn't be too large.

So FAIR has been effectively disbanded, LeCun is moving out, Wang is doing 996 and teams are hiring to fire to insulate people who need to vest their stock. How long until the company accumulates enough stress to rupture completely?

Lecun did not run almost anything at FAIR, he was basically an IC. FAIR has grown, not shrunk.

Agree with the first part

> he was basically an IC

Disagree with this part - ICs have to write code. He literally did nothing except meetings and WP posts.


Does this also run with Exo Labs' token pre-fill acceleration using DGX Spark? I.e. take 2 Sparks and 2 MacStudios and get a comparable inference speed to what 2x M5 Ultras will be able to do?

Is there any way to opt out? Or did European Comision just make a package deal without asking its populace for some reason?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: