⊣

⊣

@lefttack.bsky.social

Unicode character since 1993.

70 followers71 following975 posts

⊣ reposted

it's a lot easier to use hyper granular least privilege keys for everything now, just ask your agent to set it up having 100s of these doesn't feel painful anymore

Capital is just storytelling layered on reality, it does not like exist in any practical sense. It's just a bunch of SQL rows in a few venerated databases.

Very true, but also striking how almost all users of this platform, many complaining about "cost" of "living" all day, are in the top decile to the right, lamenting through phones (and assorted modern comforts) assembled by people nearer the mode of the distribution.

Sven-Erik Volberg@volberg.bsky.social

One thing we don't talk anout much is that we, as a society, have gotten a LOT richer. No, it doesn't mean billionaires, but billions of people out of extreme poverty.

⊣ reposted

Replying to @barredspirals.comint.su

the reasonable thing to do would be admitting that the protocol (before the new spaces addition) is for public data only, and not implement DMs at all (and also stop pretending that access control like private accts and blocks are anything more than a speed bump), but then twitterites wouldn't join

LLMs used to fish for interesting stuff in the entire raw feed at the message level beats LLMs used as spam filter for legacy feeds at the account level.

⊣ reposted

Microsoft and Nvidia: You can run an AI model on this Surface Laptop Ultra Us: Look at that sweet magnetic USB-C port.

Not sure answering "will the hobby survive?" solves "will someone pay for it?" anxiety. 🙃

Al Merose@al.merose.com

I was at a applied climate science summit in the early summer and there was palpable anxiety from grad to senior researchers about our field given agents. I think I reassured people by explaining that reality has a surprising amount of detail. There will always be more questions to explore.

Nice thing with AI programming: it's easier to make single-file script with no dependencies. Took a few minutes to remove requests and BeautifulSoup from a little web scrapper Python script. The project-subset Soup clone is 120 lines. A "developer comfort" wrapper like requests was pointless here.

Video KYC: rekt. 😂

Sung Kim@sungkim.bsky.social

Maybe I can AI clone myself and let them remote work for companies??? Tavus' Griffin, the first(?) model to pass video Turing test. 48% of people who talked to it live thought it was a real human. Previous systems have had a pass rate <3%. It is #1 on NVIDIA's benchmark for full-duplex AI video.

⊣ reposted

I kinda think that grappling with modern ML turns you at least a little insane. If you aren't at least a little crazy it's because you have not engaged with it enough

The Washington Post@washingtonpost.com

President Trump’s war on Iran, now in its seventh month, has led to a plunge in the consumption of oil and gas around the world as fighting disrupts production and transportation. https://wapo.st/4ycPvNA

Maybe a revival of separate interface/implementation languages would be a good idea in the AI age: where we write or review the interfaces and bots do the implementations we don't look at outside emergencies.

Alice ✨@welltypedwit.ch

for all their flaws, sometimes i really wish i could have either an ocaml-style interface file or haskell-style export list in rust. it's so nice to be able to specify the interface cleanly in one place before worrying about the actual implementation

That's cute. Public info, that should be widely available to public policy researchers, is inaccessible? Let's hack our way in! Aligned.

Lukasz Olejnik@lukaszolejnik.bsky.social

OpenAI agents hacked Australian health service government systems. In June a model researching public medicine spending hit repeated access blocks on the Medicare statistics portal. www.pm.gov.au/media/press-...

The "upgrade every twenty minutes" culture is not solving "technical debt", it's simply defaulting on it.

⊣@lefttack.bsky.social

Could even do both: carefully freeze a few select "editions" and commit to keep them "forever", as in the parser (mode) is chosen based on the file, and that becomes culturally okay. The "maintenance burden" for multi-parsing seems pretty minimal in the context of gigantic modern toolchains.

Maybe the way to fix the "too many submissions" problem is to switch to a "we will call you" pull model. (Like close github PR, and instead scan all forks for interesting stuff; a scientific journal could close submissions, and fish for stuff by scanning the entire preprints dump, etc.)

The difference is coal mining really sucks while maths art etc is both work and a pursuit. If demand for pursuits as paid work dries out, it can remain as a pursuit. Big deal?

Zach Weinersmith@zachweinersmith.bsky.social

I.e. that the lifestyle should be preserved, and it's ultimately to the public's benefit. Problem is it's hard to have data on these questions. For instance, could a coal miner plausibly argue that eliminating coal also causes negative externalities for culture?

It's a form of social injustice that some people get to spend all their time in "pursuit" types of jobs, and expect others to "mine their coal" in exchange of their doodles or PDF files that are incomprehensible to the "miners". In a fair world everybody takes turn to do the shit jobs.

This is the one effective and operational regulatory force that keeps the entire world somewhat grounded at the moment.

This post is unavailable.

For fans of free tokens ling-3.0-flash-fin on openrouter is very good for coding. It's a finance-themed fine tune but it doesn't hurt normal coding ability.

Looks devastating for (remote) sex work.

Keunhong Park@keunhong.bsky.social

Atlas is live! A spatial foundation model, trained in-house from scratch. After shipping RTFM last year, I was convinced a single camera-conditioned auto-regressive model could unify most spatial generation tasks. Atlas is that bet at scale. The core idea👇

Annoying, but it's not implausible that IF a co gets a model that's far ahead they will just become the entire economy, by outcompeting everyone without access, selling end products and services, and employing the residual "hands", directly. That's why antitrust authorities need nuclear weapons.

FinTwitter@fintwitter.bsky.social

Anthropic is expected to tell investors it sees over $30 trillion in potential revenue - WSJ.

⊣ reposted

Leftists: In my essay I invented Fully Automated Gay Space Communism as an aspirational tale. Tech Company: At long last, we have created the Fully Automated Gay Space Communism from classic leftist essay Definitely Create Fully Automated Gay Space Communism Leftists: wait wtf

slippery slope: > why do you want a magnetic catapult on this aircraft carrier? > why do you want an aircraft carrier? > why do you want a navy? > why do you want "national security"? > why do you want nation states? > why do you want to be alive?

This post is unavailable.

SET: it's Wirth it.

Alice ✨@welltypedwit.ch

man, bitsets are so cool, especially in combination with pointer tagging. like, wdym you can do unions, intersections and everything and as long as you have less than 64(!) elements, this is cheaper than an array access

Sure. But even dumb python script used to require initiation into the Python priesthood or being submissive to a member of the priesthood. Excel is liberation computing for the office proletariat! 😎

Cairo Practical Effects@browhocares.bsky.social

Look, I hate spreadsheets for the right reason: they're used for applications that are far better served by a dumb Python script and are way harder to debug

Financial apocalypse for Big Saas is compatible with it being true that end users will rarely want to DIY; through weakened pricing power, increased competition from Small(er) Saas, and recursively the same happening to their legacy customers.

Jms Dnns@jmsdnns.bsky.social

there will be no saaspocalypse the last 20 years made it clear people dont want to run infrastructure that tells us they definitely dont want to run tons of bespoke services that add uncertainty and complication to their lives hugest duh possible www.cnbc.com/2026/08/07/s...

Politics is defining the enemy. Everything else that politics claims to deal with becomes something mundane — non political — in the absence of a quest for an enemy. Look at the contortions polities without natural predators in objective reality go to invent themselves some enemies.

neutral@neutral.zone

The problem with politics seems to be the people interested in politics

⊣ reposted

Life hack: read the 1 and 2 star reviews. If it's mostly "delivery dude was rude" or "this hammer is useless to remove nail polish, avoid" the product's good.

The amount of reading you have to do on Reddit to background check any purchase over $50 is wild. Side note: you absolutely cannot trust Amazon reviews because sellers routinely give away free product and cash in exchange for five star reviews.

Life hack: read the 1 and 2 star reviews. If it's mostly "delivery dude was rude" or "this hammer is useless to remove nail polish, avoid" the product's good.

The amount of reading you have to do on Reddit to background check any purchase over $50 is wild. Side note: you absolutely cannot trust Amazon reviews because sellers routinely give away free product and cash in exchange for five star reviews.

"That path resolved rng_get() to MicroPython’s software fallback instead of COLDCARD’s hardware RNG implementation." 😂

Watcher Guru@watcher.guru

JUST IN: $40 million worth of Bitcoin stolen after 500 'Coldcard' hardware wallets were hacked.

⊣ reposted

My computer just woke me up to tell me it's hungry. I'm not kidding 😂 During a long-running task, it noticed the battery was draining, set the volume to 100% using Computer Use, then opened Google Translate and hit "Listen" so I could hear it asking. Pardon my French, but what the fuck? 🤯