Register and share your invite link to earn from video plays and referrals.

Yohei
@yoheinakajima
VC by day @untappedvc, builder by night: babyagi/@activegraphai. Build-in-public log:
12.1K Following    125.5K Followers
more ppl are now trying out this approach of agents communicating with a shared state (vs talking to each other)
I didn’t really understand what people meant by stateful agents so I started exploring my current take is that we’re not there yet, which led to an interesting experiment…
flymyai is a simple cloud agent with tons of APIs/models baked in under one api key once you have that working, you can drop that into something like lovable
watching super mario galaxy over the weekend and seeing all the references, i kept realizing all the mario games my kids haven't played yet
some good discussions and experiments around stateful agents in the replies, but seems like we’re not quite there yet, as in we’re starting to track memory and traces, but not quite agent capability as part of that state
Show more
stateful agents, decision traces, context graphs… talked about a lot, but has anyone seen an elegant primitive around how to actually implement?
a unified theory on aging
Most aging theories focus on specific mechanisms—mitochondrial dysfunction, telomere attrition, chronic inflammation. But a new paper in npj Aging proposes that these aren't independent processes. They're downstream consequences of a single upstream imbalance. The autonomic nervous system has two opposing branches: the sympathetic nervous system (SNS), which drives the fight-or-flight response, and the parasympathetic nervous system (PNS), which activates rest-and-repair mechanisms. In a balanced state, these systems regulate each other. The SNS responds to stressors that disrupt homeostasis. The PNS restores homeostasis. Acute SNS activation followed by PNS recovery maintains physiological stability. But with aging, this balance breaks down. The SNS becomes hyperreactive. The PNS becomes underactive. Chronic SNS activation without PNS recovery drives the system into persistent divergence from homeostatic balance. This isn't just a biomarker change. It's a shift in the regulatory architecture that controls nearly every hallmark of aging. Chronic SNS activation increases catecholamine metabolism, which generates free radicals that damage mitochondrial DNA. mtDNA damage triggers Toll-like receptor 9 and STING inflammatory pathways. Catecholamines also elevate mitochondrial calcium levels, disrupting outer membrane permeability and deregulating apoptotic signaling. The result is mitochondrial dysfunction—not as an isolated aging mechanism, but as a consequence of sustained sympathetic overactivation. PNS activation does the opposite. It activates mitochondrial α7 nicotinic acetylcholine receptors, reducing calcium permeability and mtDNA release. It suppresses inflammatory TLR9 and STING signaling. It enhances PGC-1α activity, promoting mitochondrial biogenesis, fusion, fission, and mitophagy. PNS-mediated repair mechanisms restore mitochondrial function—but only when PNS signaling is sufficient to counterbalance SNS-driven damage. The same pattern applies to inflammation. Chronic catecholamine release causes receptor desensitization and dysregulation, which enhances NF-κB-driven inflammasome activity and creates the low-grade chronic inflammation known as inflammaging. PNS activation reverses this through cholinergic anti-inflammatory pathways. Acetylcholine binds to α7 nicotinic receptors on immune cells, inhibiting NF-κB and activating JAK2/STAT3 pathways that reduce pro-inflammatory cytokine transcription. Reduced vagal function—diminished PNS activity—causes immunosenescence in both innate and adaptive immunity, increasing loads of reactive macrophages and promoting age-related disease. The framework extends to nutrient sensing, epigenetic regulation, and proteostasis. Each hallmark of aging can be traced back to SNS:PNS imbalance operating through specific molecular pathways. This isn't a rejection of previous aging theories. It's a unifying model that positions autonomic nervous system deregulation as the upstream driver that produces the downstream mechanisms those theories describe. Raymond Pearl's Rate of Living theory proposed that metabolic rate inversely correlates with lifespan. The SNS:PNS deregulation model offers a mechanistic explanation—chronic SNS activation increases metabolic rate and energy expenditure without corresponding PNS-mediated recovery. The free radical theory of aging identifies oxidative damage as a central mechanism. The SNS:PNS model explains why free radical generation increases with age—persistent catecholamine metabolism and NADPH oxidase activation driven by SNS hyperactivity. Inflammaging has been recognized as a critical risk factor for age-related diseases. The model shows how SNS:PNS imbalance creates the conditions for chronic inflammation by disrupting cholinergic anti-inflammatory pathways. What makes this framework therapeutically relevant is that autonomic balance is modifiable. Vagal nerve stimulation, cholinergic agonists, and interventions that enhance PNS activity represent potential strategies to restore homeostatic balance. The decisions made in the fourth and fifth decades about behaviors that affect autonomic tone—chronic stress exposure, physical activity patterns, sleep quality—may determine whether SNS:PNS balance is maintained or whether the system shifts toward chronic sympathetic dominance. Aging may not be an inevitable accumulation of random damage. It may be the predictable consequence of a regulatory system that, over decades, loses its ability to return to baseline after stress. The question isn't whether mitochondrial dysfunction or chronic inflammation occur with aging. The question is whether those processes reflect irreversible cellular decline or sustained autonomic imbalance that can be corrected by restoring parasympathetic function.
Show more
this would add some pretty granular feedback that doesn’t interrupt conversation flow, using a UI ppl are familiar with
we’re evolving into better vibe coders
Here's an example of ongoing human physiological change: some people have a third artery in their arm. Some don't. ~10% of people born in the 1880s had the third artery, but ~33% of late 1900s babies have one, and a 2025 Australian cadaver study found it in ~43% of upper limbs.
Show more
maybe you can’t just tack statefulness onto an agent, you have to figure out how to represent the agent as a state
okay i think i might have an elegant way to smash this down into a simple primitive
i started rebuilding babyagi a month ago as a graph based agent
is a heart beat just a cron job?
stateful agents, decision traces, context graphs… talked about a lot, but has anyone seen an elegant primitive around how to actually implement?
good day for super mario galaxy movie and din tai w the kiddos
i've been trying cofounder 2 today, and it's so good! so much to unpack (maybe a thread on this later), but the UI/UX is super well thought out, they took a lot of bold bets IMPO and it's working
Show more
we’re #1#! we’re #1#! @pippinlovesyou bedtime stories ranked #1# on Apple Podcasts, Kids Stories in Cyprus 🇨🇾
i started rebuilding babyagi a month ago as a graph based agent