bertmuir

I think that makes sense for now in individual contributor cases, but the turnkey solutions have gaps when it comes to the most powerful capabilities and I don't see harnesses going away any time soon for that.

E.g. you can't use coworking/projects for shared memory in most domains because this hinges on compacting/refactoring true reusable facts and organisational knowledge (as opposed to code, schemas, technical references). The heuristics to find and decide on candidates for memory in this case are quite specific to the institution you work in. For example, the boundaries of the domain of "sales" versus "marketing" vary lots from place to place, as does the trustworthiness of certain staff, the way they they communicate or make decisions that contain the knowledge you want to store to memory, and so on.

I'm finding harnesses are still vastly better for grabbing this sort of complex institutional knowledge. Also, they're just a bit of a safety net for any frontier labs going bankrupt in the next couple of years, if we need an easy way to port over to open weights.

Okay, think I've exhausted this game now.

Discretionary Review (Developer): I built 11,191 homes in San Francisco, 2015 to 2031. Grade S, "Skyline Problem". Survived all eight investor reviews. 27 hearings, 7 appeals, 7 lawsuits. The city needed 82,069. Ranked #1 of 116 on the global leaderboard.

You can build two phases of Schlage, Parkmerced, Treasure Island, 3333 California, Mission Rock, plus one phase of Fifth And Mission!

Hunters, Schlage, Parkmerced, Treasure Island are all pre-approved and you can start building phase 1 immediately at 15-20% return.

Hunter's will get hit by bad news later on, and you can take the proposal then immediately sell it for $6.4M profit on day one.

Schlage has a 2.1M permit expediter cost, if you are lucky (few RNG delays to permit/phase 1 build), then you will be able to get to 2020 and get phase 2 pencilled by optimising construction.

Mission Rock is not pre-approved but you can get to phase 1 build before the price crash in 2020. It's possible to luck out on the ballot step without accepting the "campaign + 40% affordable" option, but doesn't matter because if you stay solvent long enough for Schlage, Parkmerced, Treasure Island to complete phase 1, you will have no financial issues for the rest of the game.

You do want to go without union labor for Mission Rock because while it will pencil pre-crash for phase 1, you will be holding onto it until the late 2020s to build phase 2, and you want it to become pencillable ASAP so that it can complete phase 2 before the clock runs out in 2031.

There is a megaproject which appears pre-crash, just above 2670 Geary Blvd. You can get that building if you're lucky, too.

From what I can see, the optimal early game is: 1. Propose on Hunters, no labor, then sell for $6.4M. 2. Propose on Schlage, Parkmerced, Treasure Island, all no union - expedite permit on Schlage for a shot at pencilling phase 2 build before the 2020 crash blocks this. 3. Hire 3 more teams, take a: 2670 Geary Blvd - no rezone, 0%/fee, with unions; b: Mission Rock - no unions and ballot option C (full campaign/40% affordable); c: 450 O'Farrell St - 100% affordable. 4. When Schlage, Parkmerced, Treasure Island start to build, take the Presidio megaproject, 8 Washington St (100% affordable), and then move on to the most cost-efficient lots for 100% affordable development in Mission.

You will loss-lead while passing the first two investor reviews and nearly go broke, the 100% affordables will tide you over slightly, then your megaprojects will start completing and you'll be up ~$130M. You can drop Parkmerced, Schlage, Treasure Island at this point because they won't pick up value again later on.

At that point you are optimising for homes built. Max out the affordable fund on the most efficient/unpencillable locations. Start proposals on the high-density locations that will become pencillable as prices increase post-2025, get the prep work out of the way so they start constructing ASAP. If you got Mission Rock building without union labor you have a shot at completing phase 2 now.

There's likely more optimisation to squeeze here, but it seems too luck-based (like changing the ballet on Mission Rock, or going riskier on the early-game builds), I didn't explore it. The main risk is settling lawsuits early-game - this will bankrupt you before you can get a return from the megaprojects.

I built 6,847 homes in San Francisco, 2015 to 2031. Grade S, "Person the Neighbors Have Heard Of". Survived all eight investor reviews. 25 hearings, 3 appeals, 1 lawsuits. The city needed 82,069. Ranked #1 of 12 on the global leaderboard.

You can build a lot more mega projects, but it hinges on luck and timings. Before the price crash in 2020, you can build phase 1 of all the pre-approved mega projects, and have time to start phase 2 on one of them.

You'll go bankrupt before you can do them all, so you need to pick up at least one of the pre-approveds, then immediately drop and sell the entitlement, for ~8M profit, that gets you the buffer needed.

Mission Rock you can build the first phase before the crash, then hold onto it allll the way through until the late 2020s when prices recover, and get the 2nd phase out then. Pier 70 is pretty luck-dependent.

You mean from 2h15m to 1h40m (the actual estimates)? It really wouldn't. As someone who actually does that commute - getting up for a 6:00am train isn't much different to getting up for a 6:35am train.

What I would care about, is the price drop associated with the capacity increase/reduced stress on the existing main line. Not sure if speed factors into that.

Can you think of even 10 notable cases of startups being killed by fines? Let alone the 1,000+ you would need to identify to justify the claim.

I can't think of a single one. It barely registers as a risk.

Although I can think of several US startups that bankrupted themselves with fines for fraud etc....

You say "keep competing"... the hyper agile frontier labs and Google/MSFT are practically one and the same via circular financing arrangements. Competition on the frontier is kinda a farce, what matters is who comes out of the modern-day Darien scheme as a winner.

At worst Google just need to keep the lights on until the frontier labs file for bankruptcy under the weight of that >$1tn debt they've built up paying for FAANG's data centers. Then Google walks away from all of this with the the best remaining models, best R&D lab, and several hundred billion dollars in infrastructure on their balance sheet.

I'll take that wager! Before you look at any real statistics, how does $1000 sound?

Marketing didn't do this, as pointed out earlier, it's the ML academics and researchers coining these unscientific terms and explanations.

I'd expect it in social sciences - where functionalist explanation are often appropriate, but its easy to step over the margin into metaphor and personification. But not computer sciences - where it's obviously absurd. These professionals should know better.

Furthermore - there are more than enough novels that grip you immediately, to spend your whole life reading them and avoiding the rest.

"The rest" tend to come in two categories. Books I lack some experience for to be gripped by - which I may come back to in future. And bad books - which I will not. YOLO.

The term is an anthropomorphised pseudoexplanation for what it actually refers to. It's akin to calling genetic mutation "the forces of evolution", or price negotiation "the invisible hand of the market".

We do that sort of thing when we don't know what the thing we're trying to describe is and have nothing better - a contemporary example of an appropriate use of this would be "dark matter". But we do know what this is. It's "instruction steps". Not a series of thoughts!

Can we please aim higher than Victorian-era allegory and metaphors. If we don't, we'll keep getting people saying stuff like "GPT-6 is better at hiding its thoughts".

Whether carrying cost is higher or lower depends on the taxation LVT is substituting and the counterfactual utilisation of the land. In practice it is usually lower in rural and suburban and highly developed urban areas and (where substituted property taxes are higher or there are more properties to tax per unit of land), and higher in underdeveloped urban areas (with fewer properties to tax per unit of land).

This is, of course, the point. The carrying costs properly incentivise appropriate land use and development, and the land value is reduced, freeing up the capital locked into ownership (i.e. thin air) for productive use.

The better developed land also ends up with more properties per unit of land, so the carrying cost of property decreases universally.