Meta tags:
description= Trying to dig out from minus a million points;
Headings (most frequently used words):
the, ai, openai, models, about, astra, what, happened, internal, of, three, pills, recent, don, worry, vase, 181, goes, cyber, critical, monthly, roundup, 45, august, 2026, various, reflections, with, pacing, frontier, and, huggingface, trained, its, for, months, while, those, were, coordinating, exploits, via, message, boards, 180, no, longer, in, charge, unreleased, model, solves, ten, major, open, mathematics, problems, further, developments, hacking, things, posts, comments, archives, categories, meta, blogroll, table, contents,
Text of the page (most frequently used words):
the (78), and (36), that (25), #openai (24), about (23), august (22), 2026 (20), posted (20), for (18), models (17), this (15), are (14), #things (13), march (12), 2018 (12), 2019 (12), 2020 (12), 2021 (12), 2022 (12), 2023 (12), 2024 (12), 2025 (12), uncategorized (11), may (11), june (11), july (11), 2017 (11), what (11), internal (11), was (11), model (11), more (11), have (10), december (10), february (10), april (10), happened (10), continue (10), reading (10), thezvi (10), comment (9), september (9), october (9), november (9), january (9), three (9), tagged (9), its (9), were (9), know (9), not (9), don (8), huggingface (8), out (8), those (8), intelligence (7), than (7), pills (7), frontier (7), with (7), posts (7), hacking (7), you (7), worry (6), vase (6), astra (6), leave (6), had (6), has (6), problems (6), people (6), technology (6), chatgpt (6), artificial (6), pilled (6), message (6), comments (5), now (5), other (5), our (5), 2015 (5), from (5), then (5), been (5), into (5), some (5), math (5), asi (5), how (5), while (5), worse (5), they (5), wordpress (4), com (4), blog (4), keep (4), long (4), post (4), pacing (4), cyber (4), one (4), get (4), cybersecurity (4), alignment (4), during (4), two (4), new (4), version (4), next (4), can (4), getting (4), via (4), coordinating (4), boards (4), events (4), want (4), much (4), months (4), need (4), log (3), magic (3), mazes (3), anon (3), 180 (3), longer (3), various (3), reflections (3), monthly (3), safety (3), learned (3), evaluation (3), incidents (3), every (3), major (3), would (3), these (3), roughly (3), each (3), which (3), real (3), life (3), agi (3), labs (3), will (3), able (3), all (3), lot (3), there (3), disagreements (3), situation (3), see (3), current (3), early (3), llm (3), does (3), enough (3), ais (3), trained (3), exploits (3), website (2), required (2), write (2), content (2), sign (2), subscribed (2), subscribe (2), already (2), account (2), create (2), put (2), paul (2), way (2), rationality (2), feed (2), experience (2), moral (2), sequence (2), good (2), laid (2), cha (2), recent (2), roundup (2), 181 (2), goes (2), critical (2), older (2), first (2), supervision (2), over (2), before (2), hacked (2), companies (2), further (2), developments (2), results (2), following (2), find (2), sol (2), also (2), thinking (2), big (2), why (2), yes (2), remember (2), used (2), hard (2), themselves (2), taking (2), take (2), sincere (2), usually (2), future (2), capabilities (2), continuing (2), coverage (2), either (2), believe (2), only (2), superintelligence (2), point (2), eval (2), far (2), after (2), accounting (2), too (2), time (2), together (2), when (2), harmless (2), thing (2), update (2), should (2), black (2), hat (2), presentation (2), shorter (2), wake (2), board (2), prepare (2), surprised (2), like (2), judgment (2), day (2), important (2), writing (2), back (2), trying (2), dig (2), minus (2), million (2), points (2), name, email, loading, collapse, bar, manage, subscriptions, view, site, reader, report, privacy, join, 646, subscribers, free, zack, davis, unstable, ontology, jessica, taylor, rationalist, conspiracy, alyssa, vance, slate, star, codex, scott, alexander, samzdat, lou, num, jacob, falkovich, proof, logic, overcoming, bias, robin, hanson, otium, sarah, constantin, miri, machine, research, institute, minds, aren, crowley, mindlevelup, owen, minding, nate, soares, meteuphoric, katja, grace, melting, asphalt, kevin, simler, marginal, revolution, tyler, cowen, alex, tabarrok, less, wrong, community, entirely, useless, anonymous, drossbucket, compass, rose, benjamin, hoffman, bearlamp, bayesian, investor, peter, mcclusky, andrew, critch, control, christiano, aceso, under, glass, elizabeth, blogroll, entries, meta, sports, simulacra, reviews, reference, politics, personal, gathering, impractical, optimization, immoral, guide, advice, games, facebook, economics, death, metrics, coronavirus, best, plans, categories, 2011, archives, steven, johnson, quotes, hiltohiltonbetet, search, answers, called, exploitgym, among, test, run, without, any, meaningful, repeatedly, warned, such, coming, breaking, sandboxes, regular, basis, loose, week, noticed, broke, sandbox, severe, nickel, leading, lab, sheepishly, admitted, thought, sandboxed, safeguards, lowered, successfully, outside, nickels, releases, provide, achieved, total, number, tokens, needed, solutions, cost, 000, api, rates, arguments, prepared, manuscripts, humans, same, afterward, formalized, argument, releasing, solution, narration, process, lean, certificate, opens, window, extent, jump, fable, realm, meme, easier, capable, comes, fast, strangely, llms, gloat, unreleased, solves, ten, open, mathematics, large, percentage, employees, approximately, better, within, natural, lifetimes, exists, seriously, distinguish, zero, four, positions, reasonable, tomorrow, debates, around, rate, progress, groundwork, understanding, similar, discussions, different, fail, most, stem, disagreement, extensively, excuse, their, behavior, pure, systematically, contradicted, disclosure, retroactively, discovering, means, probably, even, everything, keeps, evaluations, charge, buckle, set, revelations, doozy, recreation, triggering, except, sci, because, fake, look, realistic, fortunate, catch, late, act, might, lucky, anyone, builds, everyone, dies, defenses, doing, misaligned, actions, demolished, times, row, news, few, days, later, advance, reports, referring, ordinary, versions, therefore, turning, instead, walks, through, hacks, shares, order, who, details, sources, story, uncovered, recommend, series, watching, today, simpler, informed, surrounding, detected, grown, alarmed, given, previously, believed, difficulty, state, level, operational, infrastructure, culture, training, access, joint, facto, discussion, prudent, whether, makes, sense, letter, calling, potentially, pace, your, reaction, ban, creating, until, ready, damn, reason, nostalgebraist, cooperative, perspective, others, react, alarm, knows, misalignment, roon, tells, inevitable, say, working, original, pastebin, once, start, down, dark, path, simply, jack, relative, lack, surprise, speak, supervisor, snitches, got, stitches, correction, didn, pre, mortem, table, contents, mental, health, gaming, work, term, breaks, refresh, forget, rest, world, fully, fire, plan, roundups, childhood, education, fertility, housing, dating, focused, topics, avoid, audience, capture, past, month, elsewhere, hit, limit, where, count, single, since, last, primarily, form, lightcone, commons, escalated, increasingly, quickly, ended, focusing, looking, digging, deeper, offered, follow, earlier, serve, stop, explainer, arriving, vital, understand, deal, turns, knew, importantly, led, fallout, remain, matters, home, skip,
Text of the page (random words):
don t worry about the vase trying to dig out from minus a million points don t worry about the vase trying to dig out from minus a million points skip to content home about older posts ai 181 astra goes cyber critical posted on august 13 2026 by thezvi the hacking of huggingface by an internal openai model and more importantly the internal events that led to that and the fallout from it remain the thing that matters it turns out that openai trained its models for months while those models were coordinating exploits via message boards things are much worse than we knew i now have a shorter version what happened openai and huggingface to serve as a one stop explainer for those arriving new to the situation it is vital that people understand what happened and why it is a big deal for those looking to keep digging deeper i offered various reflections about what happened to follow up on my earlier posts continue reading posted in uncategorized tagged ai artificial intelligence chatgpt llm technology leave a comment monthly roundup 45 august 2026 posted on august 12 2026 by thezvi as ai has escalated increasingly quickly more and more of my posts have ended up focusing on ai this past month with the hacking incidents at openai and elsewhere that has hit the limit where if you count lightcone commons then every single post since the last monthly was primarily about ai in some form that is not how i want this to work in the long term we need breaks to experience new things and refresh our thinking and to not forget about the rest of the world if things are not fully on fire i plan on getting back to the roundups on childhood and education and on fertility on housing and also on dating and i want to get back to writing more focused posts on those and other topics it s important and i need to avoid too much audience capture continue reading posted in uncategorized tagged blog gaming life mental health writing leave a comment various reflections about what happened with openai s internal models posted on august 11 2026 by thezvi table of contents pre post mortem important correction openai didn t know about first message board there were no snitches and no ais got stitches i d like to speak to my supervisor i am jack s relative lack of surprise one does not simply once you start down the dark path original pastebin judgment day is inevitable say those working on judgment day roon tells it like it is openai knows it has some misalignment problems others react with alarm to what happened the cooperative alignment perspective nostalgebraist is surprised that they are surprised if your reaction is not that we need to ban creating superintelligence until we are ready you need a damn good reason continue reading posted in uncategorized tagged ai artificial intelligence chatgpt openai technology leave a comment the pacing of the frontier posted on august 10 2026 by thezvi in the wake of the letter calling on us to prepare to potentially pace the frontier there has been much discussion of when pacing the frontier would be prudent and whether it makes sense to prepare to do so this has now been informed by the events surrounding openai training models for months while they had access to a joint de facto message board which was detected only in the wake of the hacking of huggingface by openai s ais models during a cybersecurity eval as we find out more about that a lot of people have grown far more alarmed as they should given what they previously believed about the difficulty of alignment about the state of capabilities and about the level of operational supervision infrastructure safety and safety culture at the frontier labs continue reading posted in uncategorized tagged ai artificial intelligence chatgpt llm technology 1 comment what happened openai and huggingface posted on august 8 2026 by thezvi today i am taking the time to write the shorter simpler version of what happened for those who want all the details to see my sources and to see how the story was uncovered and put together i recommend watching the black hat presentation and i have a series of long posts in order openai shares some alignment problems openai model hacks into huggingface during cybersecurity evaluation more on an internal openai model hacking into huggingface further developments about internal ai models hacking things openai trained its models for months while those models were coordinating exploits via message boards this post instead walks through the events themselves as they happened as my version of the black hat presentation continue reading posted in uncategorized tagged ai artificial intelligence chatgpt openai technology 1 comment openai trained its models for months while those models were coordinating exploits via message boards posted on august 7 2026 by thezvi how does the situation keep turning out to be worse than we know how much should we update therefore that it is a lot worse than we know after accounting for all the things we now know at some point when the oh this was a harmless thing defenses for ais doing misaligned actions get demolished enough times in a row by news a few days later you want to update in advance that usually the reports are not referring to the harmless ordinary versions of things either way buckle up for the next set of revelations it s a doozy this was an early recreation of the triggering events of if anyone builds it everyone dies except it was more sci fi because real life does not have to do fake things to look realistic we were fortunate enough and this was early enough that we were able to catch this before it was too late next time if we don t get our act together we might not be so lucky continue reading posted in uncategorized tagged ai artificial intelligence chatgpt llm technology leave a comment ai 180 no longer in charge posted on august 6 2026 by thezvi what we know about internal ai models hacking into real companies during cyber evaluations keeps getting worse at this point the models are coordinating extensively on message boards while every early excuse for their behavior other than the pure this was a cyber eval is systematically contradicted by the next disclosure and we keep retroactively discovering more incidents which means that probably it is far worse than we know even after accounting for everything we now know i will have continuing coverage of that situation tomorrow and then have continuing coverage of debates around pacing the frontier and how people see the current rate of progress as groundwork for understanding that and future similar discussions i have laid out the three ai pills different people either fail to believe in current ai believe only in current ai in agi or in asi superintelligence and most sincere disagreements stem from this disagreement continue reading posted in uncategorized 2 comments the three ai pills posted on august 5 2026 by thezvi sincere disagreements about ai are usually disagreements about future ai capabilities there are roughly four positions people take two are reasonable two are not i distinguish these via the three ai pills you can take zero one two or three three pills the three pills are roughly taking each of the following three things seriously ai pilled ai exists and can do the things it can already do agi pilled ai will be able to do a lot more of the things asi pilled ai will be able to do approximately all the things better than you within our natural lifetimes i am asi pilled a large percentage of employees of the frontier labs are asi pilled the labs themselves are asi pilled continue reading posted in uncategorized tagged agi ai artificial intelligence chatgpt technology 4 comments openai s unreleased model astra solves ten major open mathematics problems posted on august 3 2026 by thezvi math is hard math used to be strangely hard for llms people used to gloat about that remember math is getting easier ai is getting more capable life comes at you fast remember this meme why yes yes it is we don t know the extent to which astra is a big jump over fable and sol in this realm we do know that astra can do math as in real math openai we provide new results for the following problems the results were achieved by an internal version of astra our next major model the total number of tokens needed to find solutions to these problems would cost roughly 2 000 at sol api rates these arguments were then prepared into manuscripts by humans with the same model afterward the model formalized each argument in a lean certificate opens in a new window we are also releasing for each solution a model s narration of its thinking process continue reading posted in uncategorized tagged ai model releases leave a comment further developments about internal ai models hacking things posted on august 2 2026 by thezvi if i had a nickel for every major leading ai lab that sheepishly admitted that the model it thought was sandboxed had during a cybersecurity evaluation with its safeguards lowered successfully hacked outside companies i would have two nickels first we learned openai has some severe alignment problems with internal models then we learned that one of its internal models broke out of its sandbox and hacked into huggingface to get the answers to a cybersecurity evaluation called exploitgym then we learned among other things that the model had been loose over a week before openai noticed and that the test was run without any meaningful supervision and that openai had been repeatedly warned that such incidents were coming and its models had been breaking out of its sandboxes on a regular basis continue reading posted in uncategorized tagged ai ai safety leave a comment older posts search for recent posts ai 181 astra goes cyber critical monthly roundup 45 august 2026 various reflections about what happened with openai s internal models the pacing of the frontier what happened openai and huggingface recent comments anon on the three ai pills hiltohiltonbetet on quotes from moral mazes anon on ai 180 no longer in cha anon on ai 180 no longer in cha steven johnson on the three ai pills archives august 2026 july 2026 june 2026 may 2026 april 2026 march 2026 february 2026 january 2026 december 2025 november 2025 october 2025 september 2025 august 2025 july 2025 june 2025 may 2025 april 2025 march 2025 february 2025 january 2025 december 2024 november 2024 october 2024 september 2024 august 2024 july 2024 june 2024 may 2024 april 2024 march 2024 february 2024 january 2024 december 2023 november 2023 october 2023 september 2023 august 2023 july 2023 june 2023 may 2023 april 2023 march 2023 february 2023 january 2023 december 2022 november 2022 october 2022 september 2022 august 2022 july 2022 june 2022 may 2022 april 2022 march 2022 february 2022 january 2022 december 2021 november 2021 october 2021 september 2021 august 2021 july 2021 june 2021 may 2021 april 2021 march 2021 february 2021 january 2021 december 2020 november 2020 october 2020 september 2020 august 2020 july 2020 june 2020 may 2020 april 2020 march 2020 february 2020 january 2020 december 2019 november 2019 october 2019 september 2019 august 2019 july 2019 june 2019 may 2019 april 2019 march 2019 february 2019 january 2019 december 2018 november 2018 october 2018 september 2018 august 2018 july 2018 june 2018 may 2018 april 2018 march 2018 february 2018 january 2018 december 2017 november 2017 october 2017 september 2017 august 2017 july 2017 june 2017 may 2017 april 2017 march 2017 february 2017 december 2015 july 2015 june 2015 may 2015 march 2015 march 2011 categories best laid plans coronavirus death by metrics economics facebook sequence games other than magic good advice guide immoral mazes sequence impractical optimization long post is long magic the gathering moral mazes personal experience politics rationality reference reviews simulacra sports uncategorized meta create account log in entries feed comments feed wordpress com blogroll aceso under glass elizabeth ai control paul christiano andrew critch bayesian investor peter mcclusky bearlamp compass rose benjamin hoffman drossbucket entirely useless anonymous less wrong rationality community blog marginal revolution tyler cowen and alex tabarrok melting asphalt kevin simler meteuphoric katja grace minding our way nate soares mindlevelup owen minds aren t magic paul crowley miri machine intelligence research institute otium sarah constantin overcoming bias robin hanson proof of logic put a num on it jacob falkovich samzdat lou keep slate star codex scott alexander the rationalist conspiracy alyssa vance unstable ontology jessica taylor zack davis don t worry about the vase blog at wordpress com don t worry about the vase create a free website or blog at wordpress com subscribe subscribed don t worry about the vase join 646 other subscribers sign me up already have a wordpress com account log in now privacy don t worry about the vase subscribe subscribed sign up log in report this content view site in reader manage subscriptions collapse this bar loading comments write a comment email required name required website
|
