Mental models, 1/n
Okay, there are 40 of them, gotta start with something:
Baader-Meinhof phenomenon, or Frequency Illusion
Similar but contrary to inattentional blindness: you start seeing things you think of/search more often and think they’re more common. Hence motivated perception, the Secret, ‘Why is everyone speaking my language here arrghh’ etc etc.
P.S. One of the links I wanted to share is actually Sketchplanations — a nice visual way to ingest things 🙃
Ostrich effect
Didn’t think there’s an effect explaining my life.
Anticipating discomfort is often worse than the discomfort itself, so we avoid it altogether. Goes amazingly in line with “Choose not to be harmed — and you won’t be harmed” by one of the stoicism OGs, Marcus Aurelius.
Nobel disease
Looks like a subset of Halo effect bias, but double-sided:
- Excelling in one field doesn’t prevent you from being incompetent in others diverse feedback is crucial;
- When assessing someone’s opinion/theories, do not let their status in one field influence your feedback bullshit is always bullshit, being universal like love 🫀
Warnock’s dilemma
Especially relevant in today’s attention economy: news are often that provocative and fishing for our emotions only because it’s the only way to grasp but a tiny bit of our attention and time.
We’re spending our focus like we’re some fucking businessmen. Okay, cute fat cats are good, and cuteness is healthy:
Google Scholar effect
It’s basically the Matthew effect and Filter Bubble combined:
- Google Scholar ranks articles not in the least by their citation count cited articles are more likely to get cited again, powered by Lindy effect inverse: the more something is cited, the longer it will live as a meme of information;
- Highly-cited articles prevent the less cited/maybe ones expressing other opinions to surface, making science a bit, may I say, prone to recursion circle-jerking?
Just to think of: Google Scholar effect combined with Nobel disease can bring harmful consequences to the world’s scientific knowledge and progress.
Migraine
MoA
Migraine is a bitch, there’s no arguing that. Some of the latest theoretical mechanisms is as follows:
- Potential and optional: a wave of so-called Cortical Spreading Depression causes excitation in response to stimuli or triggers, then is changed by inhibition of neurons in the cortex. This may cause the migraine aura.
- TRPV1 receptors are expressed more in chronic sufferers’ scalp arteries and may promote activation of the trigeminovascular pathway pictured below. This pathway may be one of the main pain-signaling ones. Fun fact: TRPA1 receptors can, too, be used to promote headaches!
- TRPV receptors also promote the release of CGRP (potent vasodilator it opens/dilates blood vessels on the scalp) and Substance P (just a transmitter of pain, simple as that). Other peptides suspected to be in action are VIP (lowers pressure and dilates blood vessels) and pituitary adenylate cyclase-activating polypeptide-38 (PACAP-38) (causes mast cell degranulation/activation and vasodilation).
- CGRP and Substance P both signal pain and can lead to Central Sensitization, a state of heightened pain sensitivity due to repeated trigeminal pathway activation and inflammation.
Potential drug MoAs
- 5HT1B/D agonism: potentially inhibit the release of CGRP. There’s a wide family of so-called triptans with that MoA.
- 5HT1F agonism: also modulates CGRP release with less risk of vasoconstriction (narrowing of blood vessels). The newly approved brand-spankin’-new 5HT1F agonist is…lasmitidan!
- CGRP receptor/peptide antibodies: fremanezumab, galcanezumab and eptinezumab target the CGRP itself, while erenumab targets the CGRP receptor.
- PACAP-38 antibodies: another peptide promoting migraines. There was a trial of PAC1 (receptor for PACAP-38) monoclonal antibody, AMG-301, where it didn’t show any benefit compared to placebo.
- Onabotulinumtoxin-A is a multi-modal chronic pain inhibitor:
- Trigeminal pathway inhibition via numerous mechanisms: topiramate + I suspect other calcium channel blockers able to cross BBB (blood-brain barrier) to be effective?
- Just random thoughts: substance P antagonists? However, they didn’t show any benefit.
- Non-pharm interventions: American Headache Society proposes that stimulation of sphenopalatine region nerves promotes significant relief in migraines. They also mention that there’s only one migraine specialist per 65,000 sufferers — both a pity and an opportunity 🫣
Linkies
A disruptive product per year
…by Aakash Gupta, since 1977!
KG and NaLLM
Combining knowledge graphs, text and vectors is the heat for causal reasoning with LLMs. Neo4j is exploring that with their NaLLM.
Warpdrive and faster multi-agent RL
Heyy, let’s just transfer all the needed data into a GPU, then create CUDA blocks for environments and simulate each agent in a separate thread? Yay, releasing Warpdrive (looks actually pretty interesting for MARL)
Twitter community growth
Some semi-obvious advice from Indie Hackers:
- Identify low hanging fruit — the audience that loves your content. Return with a rage better content;
- Check for hashtags that did well use them;
- Grow engagement, be alive: comment, like, retweet, list etc;
- Start connecting with people in active communities you join;
- If returning, do an “I’m back”, then “New content” campaigns;
- Cross-post and cross-share your profiles and content where applicable and permitted.
Tax liens
A noice potential way to invest given you’re in the know and ready to diversify:
- Owner unable to pay property tax it’s now in tax lien issued by the government;
- The investor pays it off awaits return + interest from the owner;
- In the case of non-payment, investor has the right to redeem the property.
Risks
- Unknowns: you may acquire a dangerous/toxic land;
- Tenants/neighborhood: What you say about my mama ? — YouTube, self-explanatory;
- Dilapidated (Love that word. It’s like “She wants the D” someone dilapidated and barely alive) properties.