Pablo Farias

Founder of Zenva. Navigating the Ruliad at my own pace while teaching +1.5M people coding and gamedev

[Pablo Farias @pablofariasn]

@rv_sehgal the methylated results in TruHealth (e.g. vitamin C status), what time period do they (roughly) correspond to? weeks, months, years, decades? I know its hard to answer, but it's what makes or breaks the test's usefulness imo thanks in advance!

[Pablo Farias @pablofariasn]

I'm 43 and my ALP went up from 48 to ~60-70 in a year. Main changes were to increase exercise from 2 to 6 times/week, clean diet, drop alcohol and improve sleep. How would you know if ALP increase is due to bone turnover from exercise or increased bone destruction, given healthy liver markers?

[Pablo Farias @pablofariasn]

We were censored by a bunch of neo-ludites but that's not going to stop us. GameDev Assistant beta launching very soon 🚀🚀

[Pablo Farias @pablofariasn]

@sgt3v @threejs @godotengine The other factor is whether an engine's community WANTS better AI integration. With @godotengine it's dividing: our learners love AI, but the Reddit subred hates it unanimously (our last devlog was censored for showing like 5 lines of AI-generated code)

[Pablo Farias @pablofariasn]

In @godotengine you can do everything programmatically, but LLMs aren't trained enough on the API so you need to spend lots of tokens in just docs context. We are building an agent for Godot that draws game context (e.g. read scene) and do stuff (e.g. create nodes, scenes, scripts), so you can combine the best of both worlds. We are launching closed beta soon: https://t.co/xnmBoR644k

[Pablo Farias @pablofariasn]

@yeeseak @pieter_mj_j @Madam_Mito @Elroy168 @Madonna how much were you mega-dosing with? also the current 3 grams is that only EPA+DHA, or 3g of fish oil total?

[Pablo Farias @pablofariasn]

@Madam_Mito Very interesting thanks!

[Pablo Farias @pablofariasn]

@Madam_Mito what are your thoughts on thymus regeneration? This stuff: https://t.co/3y6eVKqF10

[Pablo Farias @pablofariasn]

@MatthewBerman Has anyone compared it to Mistral Small in terms of intelligence and real world usefulness?

[Pablo Farias @pablofariasn]

@Madam_Mito @InTime_49 @statto Does that mean actual anti aging is gonna be a device, not a therapy?

[Pablo Farias @pablofariasn]

@DeepInfra The model is not longer on the platform

[Pablo Farias @pablofariasn]

@LeoVasanko @abacaj Same here. I used the Hyperbolic API. Consumes more than 8k "thinking", which doesn't happen for us with R1

[Pablo Farias @pablofariasn]

@abacaj I should add my test was with @hyperbolic_labs . As other people mentioned in the comments, it could be a quantization issue

[Pablo Farias @pablofariasn]

@abacaj Hard pass for us. Tested on large contexts for our gamedev assistant. Uses huge amounts of context window thinking without ever wrapping it up for the user. Couldn't complete thinking with 8k max_tokens, whereas R1 can do that 9/10 for us.

[Pablo Farias @pablofariasn]

@togethercompute is your "Qwen/QwQ-32B" the one that just came out?

[Pablo Farias @pablofariasn]

@bryan_johnson How far can you get with just the basics (diet, sleep, exercise)? or even supplements, but nothing "experimental"?

[Pablo Farias @pablofariasn]

@MattiSchroder @Trustpilot @judgeme You don't have to pay to invite users to review in TrustPilot. We've used it for 7 years without ever paying a cent. Our rating matches exactly what you can calculate in Excel. It's a lot of work and effort to have good reviews, but worth it in the long run

[Pablo Farias @pablofariasn]

@alexalbert__ @AnthropicAI Will you provide an OpenAI compatible api endpoint?

[Pablo Farias @pablofariasn]

@EastlondonDev Wow I'm glad I'm not the only one. I've been going back and forth between 3.5 and 3.7 for some complex issues, but unlike you not seeing a real edge in 3.7 yet

[Pablo Farias @pablofariasn]

@nabeelqu o1 is way smarter than o3 in super complex problems for me at least