SWL Week in Review - Protest

From:

sam lessin <s@wlessin.com>

To:

Hidden Recipient <hidden@emailshot.io>

Date:

8/17/2025, 6:22 PM(sent yesterday)

More or Less :: AI Flashlight… or Something Else Light

If you don’t get a protest at your conference, you probably didn’t market it well enough… Hollywood’s new business model might be extras for the staged protests. Perplexity (which no one uses anymore) + Chrome would be even better than AOL-TW… but really Taylor should throw her hat in the ring with an Erez SPAC. Plus why we need a explicit conspiracy LLM & why the AI device of the future is either a flash-light, or a something-else-light…

Hot Takes

Someone needs to explicitly build and market a conspiracy generating LLM — put in a person, entity, and facts — and the LLM will invent the conspiracy for you! This is what everyone wants anyway… someone needs to just give the people what they want (or really if you insist on being high-minded, doing it explicitly will be both hilarious and hopefully help inoculate everyone … all these AI kids are desperate for a narrative to be the next Einstein, Von Neumann, and work on the Manhattan Project — how about someone stand up to be the AI Jonas Salk?

The e-cane, the AI flashlight, or the AI —light — the future of a AI hardware probably isn’t a lamp… but there is a thread to pull here…which ends where it always starts — OnlyFans style.

If your conference isn’t attracting protestors — you probably need to hire actors to keep up appearances — seriously, forget Sydney Sweeney’s Genes… if you aren’t attracting protesters in 2025 you need to figure out a better marketing strategy …

I might not like AI SaaS… but I would argue I am all AI SaaS myself — I think this is very punny… but no one else seems to enjoy my joke here… so I am going to just keep hitting it until someone acknowledges me.

Oh Harvard… — a few things.. first, it is kinda wild that as a percent of the population, Harvard’s 400K alums is actually lower than the numbers they were doing in 1700… they really should fix that. And of course settlement scuttlebutt (obviously they have to settle, Harvard is literally alone something like 50% of medical research in the country!) — Allison Wu who co-founded 1636 forum with me and runs the org in the Globe… “If Harvard settles, it’s important to do it in a way that ensures students and faculty have the stability to focus on academics, minimizing whiplash from escalations over enforcement disputes,”

Lawfare — we all know that DT is pushing this to new highs / and new level of visibility — but there really is something going on with people using lower-cost-of-litigation as a lever … especially as lawyers get MORE expensive for the average person and firm with strong return on scale/power dynamics.

The Lumpy Units of Labor You Can Purchase — related but separate. The newsletter version is ‘why is it more efficient to build a house than to renovate a bathroom’ .. and what do you do when you can only hire housekeepers full time… you end up with a pretty nicely stocked mini-fridge in your office.

“Give me the best three arguments against” — This is a Larry Summer’s line / and it is right — if you present an idea I really do want your best three reasons you are wrong along with your recommendation — and if they aren’t good it is reasonable to think you haven’t really thought it through enough. We do this to a point with investment memos ‘if I am wrong why’ — but we need to double down on taking this seriously.

Understanding Human Incentives becomes more and more critical as classical economics falters — capitalism is this incredible open protocol for working with people because all dollars are the same (no discrimination) AND the theory is everyone wants them similarly. Unfortunately this is breaking down with order-of-magnitude capitalism, the re-rise of fungible social capital & option value in an age of extremes (covid, etc.) — so it becomes really really critical to understand the real personal INCENTIVES of your counter-parties like never before (which unfortunately is both hard and time consuming) — I miss the ec10 version of the economy.

The millennial ‘Classpass’ VC subsided life party didn’t pay off… will it with GPUs? — this is the question as these AI companies spend into oblivion…. We know what happened after the millennial gut (the clean girl / can’t afford anything aesthetic) — what is the AI equivalent…

20VC Conversion — a reader pointed this out and it is funny. After getting super mad at me for pointing out that SaaS kinda sucks and most of their companies don’t matter — it seems the 20VC team is starting to internalize reality :)

Finding a place to demonstrate competence and mastery in a world of AI — great framing of what is broken — especially for young men - about their social space. Unlike women men aren’t nearly guaranteed sex and procreation…. So finding where and how to demonstrate competence and mastery and do their mating dance to pass on their genes in a globally liquid market — that is the rub…

Please Meet Gavin Newsom the National Internet Character … Personally I prefer Eric Adam’s super weird Instagram…and even Andrew Cuomo increasingly unhinged twitter (he got the assignment… just a little late?) … but really Gavin, not even the extreme liberals are calling for Garber’s resignation & this is just counter-productive pandering.

Opportunities : Requests for Projects & Jobs + Events & New Sources

[new] Slow Creator ‘CEO Summit’ Event — Team slow creator is putting this on to help creators think about how to build REAL businesses in their verticals / not just do the sponsored post thing. Everyone gets this is where the world is going, but we keep finding actual education on this shift / what to do sucks — should be great if you or someone you know wants to join (70 ppl / intimate first event)

[new] Franklin Forum for Penn — At Harvard we now have the 1636 Forum as an independent newsletter + alumni organization trying to productively support the university & focus on academic excellence as the ‘single bottom line’ everything else supports — we are now helping Penn alums do something similar there… If you are a Penn alum or know folks who might care — sign up. Content is great & important.

[new] Lettermeme — you probably know I am having a lot of fun with this — building my own ‘smarter’ view on what the intelligent world is paying attention to — if you haven’t you should sign up — while still evolving it is getting good && the cartoons the AI is now generating are IMHO a hilarious summary of the day.

[old but still] Engineers for Duolingo for Religion — still looking for folks interested in working on this - obvious market need short term around conversion (converting for my wedding / partner, need to learn this religion — just like language) — longer term lack of meaning / search for community on red alert — priority on people who can code/build.

[old but still] Entrepreneurs for LLM Output Betting Markets — I am really hot to trot on this theme / idea… the problem with prediction markets on real world events is they are too constrained and too slow for good high-frequency trading… But betting on LLM model output (on x date the answer to y query without context will be Z) — is absolute jam. High frequency ready, oracle problem solved - and actually in an interesting ‘markets are information’ sense really quite an interesting angle on model quality and accuracy. HMU on this.

[old but still] Members for Foundation Farms — I continue to be more bullish on foundation farms (our net jets for farms take to get you opted-out of the awful factory farmed food system that is killing you. Lot’s of people love the idea — and good early momentum, but still looking for more founding members to join (we find basically if you have young kids you don’t want to poison and can afford it — that is the zone)

[old but still] Childcare / Nanny for us — We need to add a new nanny to our family team in Hillsborough - full time is default but open, most hours are afternoons and a few evenings in hillsborough. Ideally flexible to travel - for more context on what this job is see the boys antics at instagram.com/lessin

Sam

P.S. Beware Nantucket Fog…as a 42 year old I understand why my dad made cape a priority so he could always drive off when he needed to. Very reasonable.

P.P.S. This is a good point— if you know with a certainty of 1 that the car will stop — traffic is gonna suck. Critical slight distrust of human drivers might be hard to get into the AI platform … at least legally (don't stop 1/10000 hard coded variable is gonna be a real legal issue)

P.P.P.S. god bless the internet.

Similar newsletters

There are other similar shared emails that you might be interested in: