Since this morning the search also reads the EPC and general contractor. It uses the same rule as the project page, so a builder only matches when the page shows that name with a source. https://t.co/6mFGB1tx7u
Searching "Bechtel" in ScopePlex returned one project. Bechtel is the named EPC on 10. The search read project names, owners, places, design firms and permit numbers, and never the builder.
@rtehrani@torusrev A prefabricated data center is still a fabrication job, just done in a shop instead of on site. I'd be curious where Torus builds the modules.
@bofrench That list will be a useful record in itself. Big sites with their own generation also leave a trail in TCEQ air permits, which is one of the places new Texas data center campuses show up early.
@eliana_jordan Number 3 is underrated. One member test on my app found five pages counting projects five different ways, and my own clicking had never caught it.
@itsharmanjot I keep the context in the repo instead of the chat. Every pull request on my project says what changed and why in plain words, so a fresh session reads the history instead of guessing.
@PrateekJainDev Local testing gets me most of the way. The bugs that reach my users usually come from the data, like a search that never read the builder's name, and no emulator catches that.
@Bic_Revelation This is the risk I watch most. Agents write most of my code, so every pull request has to say in plain words what changed, and I check it on the live site before it counts as shipped.
@sysxplore A file like that is only as good as how specific it is. On my own project the rules that hold up are the ones written after a real mistake, like a QA check on the live site before anything counts as shipped.
@AntonMartyniuk Code review in the same window is the piece I'd want. I read every change next to what the live page shows, so less switching between tools helps.
@IlirAliu_ Reusing the old arm is the smart part. I build my app with Claude Code too, and the work that takes longest is checking the output against the real world, which for me means real permit filings.
@KareemFarid A gas plant that skips the grid queue still needs an air permit. In Texas that usually means a TCEQ filing, which is where I'd look for the Starbase plant first.
@aaliya_va Trust is the part I can't generate either. Every project record in my app cites the public filing it came from, because a reader in construction won't act on a number they can't trace.
So every project now gets a type from its filings: new build, expansion or upgrade. A running plant that only files permit upkeep leaves the site, and one with new work, like an LNG terminal with a fresh incentive application, stays as an expansion. https://t.co/6mFGB1u4X2
A permit filing doesn't always mean something new is being built. Running plants file renewals for years, and ScopePlex was showing some of those plants as fresh leads.
@akshay_pachaar What bit me in production wasn't the agent's reasoning, it was the data it touched. One bad row in a batch of public filings used to sink the whole batch, until failed batches started retrying row by row.
@TheValueist The read-through I watch is the Louisiana paper trail. Incentive applications and air permits can show up well before the equipment orders get discussed on a call.
@suraj_sharma14 Spec writing is the one I'd move to the top. With agents writing most of my code, a vague spec gets built exactly as written, and the bugs that follow are mine.
@china_saparbek Respect for posting the minus next to MRR too. My build log is much younger, and so far the fixes users notice most are small ones, like a project name reading like a name instead of a filing title.
@mirdhaaakanksha Public data on a map is underrated. I do something similar with Texas and Louisiana permit filings, and the hard part isn't plotting them, it's deciding which records are the same project under different names.