Goderguy

making drugs & havin fun. prev L/S analyst

[Goderguy @Goderguy]

“loops” is the word of the day post ANT model. what is done in code can eventually generalize to real world experiments.

please reach out if you are building these to share notes

[Goderguy @Goderguy]

value to a biotech org around designing these loops as a forcing function for program development are underpriced imo

limiting constraint on any biotech today needs to be the speed of assay biology. goal is to set up infrastructure to hill climb against what are fundamentally out of distribution outputs: affinity, potency, etc.

this will quickly transition to a world where experimental outputs are reasoned on by agents, new sequence modifications or recommendations are made autonomously, and data is collected + reasoned on 24/7, in an effort to improve a discrete pre-defined metric, while minimizing risks to downstream liabilities in lead op

[Goderguy @Goderguy]

value to a biotech org around designing these loops as a forcing function for program development are underpriced imo

limiting constraint on any biotech today needs to be the speed of assay biology. goal is to set up infrastructure to hill climb against what are fundamentally out of distribution outputs: affinity, potency, etc.

this will quickly transition to a world where experimental outputs are reasoned on by agents, new sequence modifications or recommendations are made autonomously, and data is collected + reasoned on 24/7, in an effort to improve a discrete pre-defined metric, while minimizing risks to downstream liabilities in lead op

[delian @zebulgar]

@SinaHartung @jacobkimmel I’ve looked pretty broadly at this category.

The differentiated thing isn’t just “longevity assets.”

It’s that they've done full loop. Models of cell/organ age → large-scale mRNA TF payload screening → functional assays showing rejuvenation while preserving cell identity

[Goderguy @Goderguy]

@SimonDBarnett @plainyogurt21 haha missed yall and our online relationships

[Goderguy @Goderguy]

with the right custom harness - # of discovery campaigns i can reliably prosecute has increased from ~3 to ~7-8

very similar to stock coverage - there is a high up front “reasoning cost” required to think through all program decision trees - this gets amortized over a large volume of “maintenance work.” most of the latter should go away

[Goderguy @Goderguy]

not just in coding

every highly specialized unit of labor not interacting with the physical world can be productized with the right harness + reasoning model architecture.

agree these assets will trend towards oligopolies in every domain - only issue remaining is the non-verifiable nature of most knowledge work. will create opp'ties to compete vs. labs imo

[Goderguy @Goderguy]

@SimonDBarnett congrats sir - amazing seeing you execute this

[Goderguy @Goderguy]

@midwit_capital @liteupthedark1 you're the man!

[Goderguy @Goderguy]

@liteupthedark1 @midwit_capital Can I DM?

[Goderguy @Goderguy]

@Biohazard3737 empirically it’s rapidly accelerating new drug idea gen. as the wet lab workflows are commoditizing & lead op moving to computation.

[Goderguy @Goderguy]

@TKSaville @broheim777 Thanks for all this incredible detail. Who do you view as the number two player in this base? Seems outside of Comstock there are multiple sub-scaled players - but tough to tease out who else has a “right to win”

[Goderguy @Goderguy]

biggest misconception I see from $LODE: the company is booking a net income accounting loss given their majority ownership in Bioleum.

But - this is not a cash drag on $LODE equity given it's independently financed. Worst case - PF equity ownership to $LODE shareholders in Bioleum goes down if they do a raise. The accounting loss is not the right representation for the biz's cash flows.

[Goderguy @Goderguy]

@biocheMichael Amen Michael

[Goderguy @Goderguy]

@DdelAlamo More for VHH’s vs mini proteins - but combining somewhat effective starting points with OrthoRep or alternative YSD platforms seems like a higher confidence approach? Curious if you’d disagree.

[Goderguy @Goderguy]

@SimonDBarnett @deepseek_ai this was great

[Goderguy @Goderguy]

@0xsmac @notncssry do you think ridges or the prop network are steps in the right direction? Or not feasible

[Goderguy @Goderguy]

@Respekchemistry nah I agree with you - KCC2 is too risky for now. their other asset - OV329 - with data Q126 is what I was referring to

[Goderguy @Goderguy]

@Respekchemistry buy on this dip into Q126 OV329

[Goderguy @Goderguy]

what do you think the solution here is? if reserves are removed as regulatory leverage for SLR - banks can incentivize deposits+reserves and thus their ability to provision liquidity without tightening markets.

Not sure what else we have. Seems Fed governors are okay with waiting until something breaks before moving.

[Goderguy @Goderguy]

@biocheMichael How are you assessing binding quality here?