b0a04gl

17 ms·

61.
by b0a04gl 1y ago
the ai agent runs inside the same remote runtime as the app. does it share the BEAM vm or run as a port process?
62.
by b0a04gl 1y ago
rtc handling 300ms+ delay and still pulling off tasks like plugging ethernet is kinda nuts. what i'm not getting is but how's it keeping the control loop stable without retraining? some sort of latent plan caching?
63.
by b0a04gl 1y ago
if they trained this scale without western cloud infra, i'd want to know what their token throughput setup looks like
64.
by b0a04gl 1y ago
yes im agi by the way
65.
by b0a04gl 1y ago
96.8% accuracy sounds impressive until you realise they skipped the REM phase like it's a bug report. "user unpredictable when dreaming, exclude from dataset." also love how your breath is now a biometric. imagine getting loc
66.
by b0a04gl 1y ago
when o3 pricing dropped 80%, most wrote the entire model family off as a downgrade (including me). but usage patterns flipped people finally ran real tasks through it. it's one of the few that holds state across fragmented prompts with
67.
by b0a04gl 1y ago
my older brother used to dial into local BBSes late at night, tying up the phone line and pissing off everyone. mostly forums, file sharing, a few ascii games. he showed me how to telnet into some later on when it moved online. that story a
68.
by b0a04gl 1y ago
yeah fair, i get that not everything has to be polished or production grade. and your work has that clarity purposefully scoped, honest about tradeoffs. that's rare.. i'm all for hobby tools and experiments, just want us to be upf
69.
by b0a04gl 1y ago
everyone saw this coming the day facebook bought it, but the real issue isn't ads in status . it's that the platform is now locked into meta's attention monetization engine. the founders explicitly said no ads. now not only a
70.
by b0a04gl 1y ago
these get into logs, metrics, API contracts. now you're explaining isUserAGoat() to a partner team in a quarterly review. nobody’s laughing then funny thing is, linters catch unused vars but not unfunny ones. maybe we need a linter tha
71.
by b0a04gl 1y ago
on a recent trip, i stayed at a hotel that had a 1-hour free wifi trial per device. after it expired, i toggled mac randomisation and reconnected. worked again. did it four times, no issues. never paid. later checked the captive portal doma
72.
by b0a04gl 1y ago
how do you handle cases where the seek table itself gets truncated or corrupted? do you fallback to scanning for frame boundaries or just error out? wondering if there's room to embed a minimal redundant index at the tail too for safet
73.
by b0a04gl 1y ago
what if prison ends up becoming the most distraction-free dev environment. no meetings, no slack pings, no linkedin recruiters, just you, a terminal, and 10 years of uninterrupted focus. kinda terrifying how productive that sounds
74.
Don't Build Multi-Agents(cognition.ai)
4 points by b0a04gl 1y ago|1 comments
75.
by b0a04gl 1y ago
but who has the gear, who keeps it charged, who actually shows up when the net goes dark. tech's the easy part. the hard part is getting 5 neighbors to agree on a channel, a meeting point, and a backup plan they’ll actually remember. a
76.
by b0a04gl 1y ago
any channel that hides data invites misuse, even if the intent is benign. the same technique that enables structure also enables stealth. how easy it becomes to slip data past casual inspection. worth being cautious, even with open use case
77.
by b0a04gl 1y ago
lisp's power came from its default being openness. not just macros or homoiconicity, but how the system assumed you would reshape it. that expectation isn't present in modern languages. today, customisation is an advanced feature,
78.
by b0a04gl 1y ago
using LSB to store structured metadata inside PNG is clever > survives format conversion, stays invisible to standard viewers, and doesn't break compression. but the space is tight. even at 1 bit per channel, that's just 3 bits
79.
by b0a04gl 1y ago
got it, thanks. makes sense that it depends on the projection target. SHACL+SPARQL seems like the strongest runtime check layer then. for projections like graphql or java where enforcement is weaker, is there any way to inject runtime guard
80.
by b0a04gl 1y ago
how much of upper is actually enforced at runtime vs just used for schema generation? like if a downstream system silently breaks a semantic assumption (say, infers enum incorrectly or drops a type constraint), does uda catch that anywhere
81.
by b0a04gl 1y ago
i tried writing a tiny logic engine once after reading the SICP chapter and yeah, the syntax part was easy to mimic but the actual backtracking logic hit me harder than expected. what helped was thinking of it less like solving and more lik
82.
by b0a04gl 1y ago
snake at 1fps is probably the most cursed form of patience training i’ve seen.
83.
by b0a04gl 1y ago
yeah i got what it was going for eventually, but tbh it was annoying at first. the scroll interaction wasn’t clear and it broke the reading flow. felt more like a bug than a feature until i slowed down and figured it out. the context jumps
84.
by b0a04gl 1y ago
>She lives below luck-level, never imagining some lottery will change her load of pottery to wings. nails the mindset where imagining change doesn’t even happen. it’s not about failing to win. it’s about never thinking you’re in the draw
85.
by b0a04gl 1y ago
fair, recursive groups exist, and yeah stage 2’s structure is solid. but the point was less about recursion as a feature and more about context awareness. vim’s engine lets you nest, sure, but it doesnt preserve intent across transitions. y
86.
by b0a04gl 1y ago
vim’s syntax engine doesn’t track context. it matches tokens, not structure. in langs like ysh where command and expression modes mix mid-line, this breaks. no memory of nesting, no awareness of why you’re in a mode. one bad match and sync
87.
by b0a04gl 1y ago
i gave it a 4 step research task with branching subtasks. told it upfront what the goal was. halfway through it forgot why it was doing step 2. asked it to summarise progress so far and it hallucinated a step i never mentioned. restarted fr
88.
by b0a04gl 1y ago
> "Finally got sick of diffing contracts in Word and emailing PDFs back and forth" this is the actual iceberg. forget ai clause generators and buzzwords, 99 percent of the pain is version chaos. localfirst editor is cool and im
89.
by b0a04gl 1y ago
true, you want microsecond decisions at the core, no doubt ; but that’s only half the game. an ideal action in clean memory isn’t the same once it hits fragmented liquidity, stale quotes, partial fills. if the algo doesn’t account for execu
90.
by b0a04gl 1y ago
you mentioned SMT is slower than SAT and left it there, but that feels incomplete. in problems like this, solve time barely matters unless you’re generating at scale. the real weight is in how fast you can write, refactor, and trust the c

More ›