> Because I don't have a garage or a driveway I can't park right at my house, so I would not be able to charge at home.
IMO, this matters more than anything else.
If you can't charge at home, then an EV becomes rather nonviable unless you drive very little and your usual grocery store has a fast charger. But then public chargers are typically at least double the price per kilowatt hour compared to charging at home.
That's not a concern it's a reality. Iran is not shut-off or blockaded to any meaningful degree. It has tons of unmolested border crossing and Caspian sea access, and maintains full control within it's own borders (minus the parts that have been blown up).
Similar effort with PageIndex [1], which basically creates a table of contents like tree. Then an LLM traverses the tree to figure out which chunks are relevant for the context in the prompt.
I strongly dislike flexible input like
__Unambiguous___, *Unambiguous*
I’m reminded of the time Microsoft allowed mistakes in html writing. They attempted to parse a wide variety of common user errors. The effect of this was no standard and nobody else able to write a Microsoft compatible parser.
I dislike Nim lang because of this. At least Nim defined the specification. Still though I think it creates more cognitive load learning every legal variation and it makes searching more difficult.
I think to authors point if Markdown actually had a strict simple definition with one day to do it and no embedded html we would be better off.
In undergrad I had a buddy who was a political science major, and he put it pretty bluntly one day: "Do you engineers realize how arrogant you sound when you're talking about things you have no clue about?" 20 year old me just laughed and thought to myself "lol liberal arts majors" but now that I'm older and more grown up, he was totally right and I see it all around me in the industry. Especially here on HN.
A genuinely good-for-the-world project. The data is really useful for science and for machine learning. You can export all the research-grade identifications of fungi to train a classifier; if that’s what you’re into.
JS numbers behave much more like C's definition of signed overflow being UB as it's signed numbers are effectively like 51-ish bit with a SEPARATE sign bit and non-assiciative behavior when overflow happens.
> A second Air Force combat plane crashed in the Persian Gulf region on Friday, and the lone pilot was safely rescued, according to two U.S. officials who spoke on condition of anonymity to discuss operational matters. The A-10 Warthog attack plane went down near the Strait of Hormuz about the same time that an Air Force F-15E was shot down over Iran, the officials said. In that incident, one crew member was rescued and search-and-rescue operators are looking for the second airman. Officials provided scant details about the A-10 crash, including how and where it happened.
The platform I built is live in beta at FluentLogic.org, for real families. I’m a high school teacher with a physics and philosophy background (no software engineering experience) who spent 10 months building it—roughly 350,000 lines of production TypeScript, written entirely with AI assistance. I don't know TS from JS. I know assembler and C++.
No matter how many times I asked the model to audit the same piece of code, I kept finding the same categories—until I forced a completely different angle. New class of bugs. Then a plateau. New angle. New class. Plateau again.
Before this, I tried the obvious: firing hundreds of varied prompts, changing phrasing, and hoping coverage would emerge from volume. I spent hundreds of dollars on this shotgun approach. It doesn’t work. You’re just sampling the same semantic neighborhood from slightly different entry points. Shotgun auditing is same-axis repetition with extra noise.
The fix is almost embarrassingly simple: add one word to your audit prompts—"orthogonal."
Instead of:
"Find bugs in this code" [or any target surface]
Try:
"Audit this surface from the most orthogonal direction to what you just found." (Then fix the bugs, rotate the axis, and repeat until you hit the P2 floor).
The models aren't broken. When you ask the same model that generated your code to audit it, you’re sending the auditor back into the same semantic compression manifold the generator already exhausted. Same manifold = same blind spot. I call this Generator-Auditor Symmetry (GAS).
"Orthogonal" routes the model through a genuinely different neighborhood, producing non-overlapping findings consistently.
The P2 Floor: When your false-positive rate crosses ~40% on two consecutive fresh-axis waves with zero new critical bugs, the surface is clean. The FP rate acts as an entropy meter.
Rotation > Diversity: Rotating a single model across 3 orthogonal axes outperformed using 3 different models on the same axis.
The Scale of the Test:
Earlier this week, I ran a 36-hour marathon audit across 150+ surfaces.
Cost: ~$10–$15 in API tokens.
Yield: 60+ P0 bugs fixed and ~150 P1 bugs catalogued. (e.g., OAuth sentinel bypasses, silent cache-invalidation race conditions). Each was invisible to other probe axes. And the web app is now feeling the snappiest it's ever been.
Same-axis repetition plateaus at ~20% bug-class discovery yield, while orthogonal rotation hits ~80% (a 4–5× advantage). I took the full 350K-line codebase to systemic P2 floor. The app is perceptibly faster afterward.
I wrote a short paper formalizing the method and the supporting topological observations. To check this wasn’t just a prompting trick, I ran persistent homology (Vietoris-Rips on Gemini semantic embeddings of 58 production bug classes). It revealed 20 significant β₁ interior loops—evidence that the bug classes form a geometric structure in semantic space that same-axis probing structurally cannot exhaust.
You can disable GC in tinygo, so if you allocate all the necessary buffers beforehand it can have good performance with real-time characteristics. If you _need_ dynamic memory allocation then no, because you need GC it can't provide realtime guarantees.
Why does being so deep in the gravity well pose an issue? If you are assuming the Dyson swarm is intended to go back up the well then sure, but that isn't necessary.
The problem is corn requires a lot of fossil fuel energy input, mostly in the form of fertiliser. The net energy output is only around 1.3 so an acre of corn produces maybe 400 gallons of gasoline equivalent output requires 300 gallons of gasoline equivalent in energy inputs.
Ethanol from sugarcane makes a lot more sense. Corn ethanol is just a wasteful subsidy for farmers paid for by drivers.
Maybe not, but I'd want to know beforehand either way. And looking through accounts near me suggests that a fair number of users add enough detail to make me think that they don't realize that their info is so public (selfies/profile pictures being the most problematic example imo).
Backed stablecoins aren't some anarchistic anti-government thing; they are highly regulated and will lose access to their banking if they don't follow the rules – rules which require them to freeze coins in cases of crime.
If you want to show a middle finger to government there are cryptocurrencies for that, but USD stablecoins with centralized backing is not it.
Timeouts are killer too... many modern apps assume slow or no response for more than 10 or 20 seconds is "Internet is down" and will stop trying.
It's brutal, even with modern Internet, when people develop apps with the assumption of the Internet connection always having decent latency and bandwidth.
IMO, this matters more than anything else.
If you can't charge at home, then an EV becomes rather nonviable unless you drive very little and your usual grocery store has a fast charger. But then public chargers are typically at least double the price per kilowatt hour compared to charging at home.