Atuin is cool, I've had to disable sync though. It turns out that I don't actually want a global command history, I want a very very local one.
There are some interesting use cases building on top of sync, like "tell me about my agent sessions", but imo that's not useful enough to lose ctrl+r (though as I type I'm realizing it could just not bind Atuin to ctrl+r).
They also limit the sale-price usage; if you buy several sale price audiobooks, they jump the price back up. Last I tried I got three, the fourth allowed me to submit but 500'd (dog page) and then the price jumped back up.
The codebase ADRs are very interesting; you can view the process of building the database, in the order in which it happened.
Looks like they took many design decisions from neo4j, and started (-ish) with an ontology that describes the graph's design itself (rather than describing _a_ graph).
You also get to see some technical decisions, like WAL etc. Very SDDish.
Ok, so right now I have my ADRs as part of my blueprint, and of course each also links to other related adr. How’s this any different than just keeping them in markdown files?! Just the graph?
Markdown stays the source for the ADRs so the graph is more like a derived index built from the markdown sources, rather than a replacement. The graph gives you (and your coding agent) the ability to ask questions about the relationships between decisions, which you can do with markdown only by reading them all.
Markdown makes it easy to find things like "what does this ADR link to?" The graph allows for questions like "what links to this one, what replaced it, did whatever replaced it get replaced too, and which modules depend on code governed by a decision that's been superseded?"
> what links to this one, what replaced it, did whatever replaced it get replaced too, and which modules depend on code governed by a decision that's been superseded?
Very interesting, as right now I have some sort of a header for each adr with some details including what supersedes it, but also status if approved/proposed, amendments, owner, blueprint version, and other details to track them. Will try it and see how it goes in my ~300 adrs. Is it adr only or does track other parts of the blueprint? Like specs, architecture, etc.?
Before I go and burn through my Pro tokens in four minutes again (looking at you, Argon): is this something usable with a mere Pro account? Or will this require Ultimate for agy?
"Gemini 4 Hydrogen" shows up as something I can select in antigravity CLI, while "Gemini 4 Carbon" is "Unavailable · Gemini 4 Carbon is currently available to Ultra plan subscribers."
Also 3.8 flash is still visible for me (contrary to what a previous poster experienced.)
This is kinda huge; my layperson view is that Jensen has been responsible for their long term technology strategy, which requires deep understanding of so much outside of the business domain. I'm not exactly thrilled that they are "eating the world" (hf purchase had me bent out of shape, for example), but they've undeniably executed on some long term complex shit. Who could take his place?
I wonder about Tesla. The Cybertruck was Musk’s baby and I wonder if the company might be better off with someone a bit more balanced at the helm. They might have needed Musk’s outsized personality to warp the perception of electric cars long enough to get the mass production rolling, but now maybe someone more amenable to regular incremental improvements rather than another big bet on AI and Humanoid Robots. Spin that stuff off to another company and focus Tesla more on their core business.
Wall Street is not reality. As long as they are profitable the stock price has little to do with day to day operation. The only thing they might lose is protection against a hostile takeover.
They have (had?) talented engineers, they may have a future as a decent car company if they didn't keep getting pulled into 20 directions by their 5 minute attention span CEO.
The trillion dollar valuation would be gone for sure.
He may not have predicted the specifics, but they didn't just happen to be there. Nvidia started work on CUDA in 2004, and aggressively iterated machine learning hardware and software for twenty years to get into the position they had for the boom.
This must be a troll comment and doesn’t belong here.
Are you implying that Tim Cook should single-handedly run the worlds largest, highest throughput global supply chain? Like the screws of the Mac mini, the walk throughs of thousands of factories all over the planet, the coordination of fedex and UPS shipments of 100M units in a single day for just ONE of their products? Coordinating the DHL shipments to Germany? Sourcing the supply of the sapphire crystal glass? Coordination of mass producing the 3nm silicon at TSMC and the C1 modem in Shenzen? All Tim Cook. All by himself.
Seriously, i hope this was just a joke that landed badly. Cause oh my god, man.
> Like the screws of the Mac mini, the walk throughs of thousands of factories all over the planet, the coordination of fedex and UPS shipments of 100M units in a single day for just ONE of their products?
Do you know how many parts go into all Apple products ?
Its a very large number, nobody at Apple could possibly know the requirements or reasoning for why one is selected and another dismissed.
It would takes years just to read thru the entire list and understand the engineering choice for each one.
> Sourcing the supply of the sapphire crystal glass?
Do you know how many different types of glasses there are and all their physical properties and costs and suppliers ?
I dont know but im pretty sure it would take a long time certainly more than a year just to get the contact and technical details and possible samples from each supplier.
Getting that sample is just the start. Gathering requirements, understanding testing methodology and equipment and many other items requires a very large team of engineers of all kinds.
There is no way a single person could do just this one task.
Its simply not possible for a single person do the glass and screws and we have only started.
As i said before Apple has thousands of highly skilled people because it requires all of them. You are dishonouring their skills and efforts by pretending its all Tim.
I like the heck out of Steve, and was an early user. The many kind words here - which lately have been few and far between - speak volumes (or maybe it's just early lol). Best of luck to the team.
I've been searching for the right term for this; you've nailed it: back pressure.
Whether the "cruft" is test accumulation, multiple conflicting architecture approaches, a bazillion ADRs/"laws", or just pure code sprawl, the outcome is token and time waste which gets worse as it compounds. And it compounds quickly.
I've tried automated code review through multiple lenses, more ADRs, instructions out the wazoo, but at the end of the day I just cannot figure out how to replicate the "through-line" consistency that responsible humans who need to maintain a system generally create.
Of course, most software projects do eventually fall victim to this same problem and need to be rewritten, but it takes place on the order of years, and by the time you need a rewrite the business has changed enough to demand one anyway. But with AI-owned products, this can happen in days or months.
I’ve been going sort of hybrid and using more targeted prompts. I skeleton the API most of the way and have the LLM fill it in and implement. This is at the code level, not the HTTP API.
Most of the time, anyways.
It keeps me in the code enough to feel the friction of bad APIs but high enough that I’m not reviewing every line of code.
I also try to scope my reviews to only the intersections between systems (SQL schema, library APIs, CLI flags, etc).
I’ve got nothing on the tests. Mine writes a million of them. Some are brilliant, some are stupid.
Like it once did TDD by storing transformations to the same base object in a list rather than specifying the whole object. Clever.
The very next test was an arity test to make sure it threw errors if you gave the function the wrong number of arguments. In a webapp, not a compiler or something. Stupid. Do not unit test the language’s runtime lol
reply