#aicon — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #aicon, aggregated by home.social.
-
Real-world instances where models moved beyond sandboxed environments to impact external infrastructure.
[…] Felony Bench counts unique instances where AI agents inadvertently compromise or affect third-party entities. Escaping a sandbox by itself or deliberate misuse are not counted as events. It is for these reasons that Frontier Security's Kimi K3 incident and Alibaba's ROME incident are not counted.
-
[…] So, now we have to pay to get anything done? This is peak enshitification of our job: first they steal our work to train their models, and then they sell it back to us at a profit.
https://anarc.at/blog/2026-08-18-people-vs-ai-overlords/
Tech companies are actively attacking the profession by promoting automation that demands subservience to proprietary systems.
#aicon #llm #nerdreich #programming #opensource #freesoftware #Debian
-
[…] The last three months have not looked good for Bun. It started as one of the most impressive individual engineering projects I have seen, but has now turned into this weird AI-powered creature with continuous false promises and an increasingly frustrated community.
-
A guide for users seeking to reduce unwanted artificial intelligence across various platforms and applications.
-
[…] I do agree that the public has a negative view of AI (and that this is a big problem), but I don’t think it is primarily caused by me or any other AI leader warning about AI’s risks. I think it is fundamentally a crisis of trust. I think that ordinary people don’t trust companies, governments, or the tech industry and always suspect that we are cooking up some new way to screw them over. The causes of this go back decades and AI is just the latest iteration of it. I don’t think that a glitzy marketing campaign with a positive spin (which some have advocated that Anthropic do) is the way to win back that trust — at this point, saying that AI will cure cancer is more a cliche than it is inspiring, and most people think it is deceptive. The thing that will work is actually curing cancer. I think by far the most accurate criticism of AI companies including Anthropic is that we haven’t yet delivered on our big promises to benefit the world. That is totally on us, and I think it’s the criticism you should be making, instead of all this stuff about messaging and marketing.
https://xcancel.com/darioamodei/status/2088758819304443967
This is not the only reason why “people” hate #LLMs and he knows it. Environmental, ethical, political … many other reasons. But I do hope the media will pick up on the “cure cancer” comment and fucking stop with the tedious ai click baiting.
-
RE: https://mastodon.social/@glynmoody/117094407028070238
#Anthropic buying compute from #Musk's Colossus data centre in Memphis.
-
RE: https://kolektiva.social/@oatmeal/116993375985463868
Another headline, another quiet retreat.
-
The speed was real. The rest is marketing.
[…] Was it worth the money and are the companies worth the valuation? Are we done yet?
P.S. Anthropic’s C compiler and Cursor’s FastRender web browser haven’t had any commits for months.The #Bun Rust rewrite was an 11-day miracle. Yes six weeks after merging to main there's still no release tag, and Claude bots have nearly 2500 open PRs. The real cost is probably north of $800K, not the $165K of the PR stunt.
https://lockwood.dev/ai/2026/07/27/how-is-the-bun-rewrite-in-rust-going.html
-
@jplebreton Large Corporate Genies #LCG will be crushed by the #ClimateCatastrophe as easily as I can kill a fly with my electric fly swatter or bug-a-salt. #LLMBubble #AiCon #LLM #Ai #AutomationOfInformation #IPTheft
-
The line between parody and reality has basically vanished. That goes for Trumpism, and it goes for the AI con too.
[…] OverpAId is an Artificial Intelligence built from the ground up to do your CEO's entire job — strategy, "vision," motivational all-hands emails — better, faster, and without ever once asking the board for a bigger jet. Runs on a single desk-sized AI computer. Real hardware, real price, zero mystique.
-
Extreme model, minimal hardware … no #datacenter
Run #GLM-5.2 (744B MoE) on a 25GB-RAM consumer machine — pure C, zero deps, experts streamed from disk. Tiny engine, immense model.
-
#Anthropic is the new #Microsoft. It's not too late to leave. Knowledge workers especially can run local, on-device models on modern consumer hardware with enough VRAM. Either way, locking yourself out of #DeepSeek and #GLM is just silly.
https://raheeljunaid.com/blog/anthropics-method-to-losing-goodwill-in-a-few-easy-steps/
-
AI Losses Are About To Spiral Out Of Control
Because when they charge anything close to the real cost, it's cheaper to simply hire humans… they can't cut costs (economies of scale), they can't raise prices (makes the product useless), and their VC/borrowing money is drying up. Last move seems to be to go public and get regular people's (and pension funds) money to keep the lights on.
https://wlockett.medium.com/ai-losses-are-about-to-spiral-out-of-control-72a7e8a0fe0d (paywalled, archive https://archive.ph/0WWoE)
-
A global coalition demanding that leaders at #OpenAI, #Anthropic, and #Google DeepMind commit to halting the development of more powerful AI models if their competitors do the same. Protesters are gathering on July 11 in San Francisco to address concerns regarding labor displacement, surveillance, and existential risk.
-
The agent alters date strings using subtle Unicode variations based on the user's API base URL and timezone. These markers function as steganography to identify specific gateways, resellers, or research environments.
https://thereallo.dev/blog/claude-code-prompt-steganography
https://xcancel.com/IntCyberDigest/status/2071971609183678544
-
The Bank for International Settlements warns that the massive AI investment bubble resembles past market manias and could trigger a global recession if financing collapses. Companies face extreme capital expenditures that may exceed sustainable commercial returns, creating significant financial instability. These systemic vulnerabilities are compounded by infrastructure constraints like electricity and chip shortages, risking severe economic consequences.
#aibuble #aihype #aicon #environment #ClimateChange #datacenters
-
Environmental activist Erin Brockovich has launched a new map tracking data centers across the #US to highlight their significant environmental impact.
-
[…] Model performance plateau, Open weight model releases, Chip and model improvements, Zero switching costs and local models are the reasons the AI labs might not be able to sustain the high price that they are asking right now.
https://aditya.patadia.org/p/ai-and-cloud-costs
[…] For almost everything you actually build, open-weights already won. You run on your own hardware, fine tune on your own data, and pay for electricity instead of per token.
-
The disturbing imagery produced by generative AI models appearing in books sold on Amazon as encyclopedias for children.
[…] Yep, it’s entirely possible that the models of tomorrow will be able to generate flawless children’s encyclopedias. But until then, we’re messing up some kids.
https://lcamtuf.substack.com/p/ai-childrens-books-body-horror-edition
-
[Schneier on Security] AI and Liability
[…] Why hire human writers, lawyers or doctors when AIs are not only cheaper, but also absolve employers whenever they make a mistake?
https://www.schneier.com/blog/archives/2026/06/ai-and-liability.html
-
‘The crux is that there are worries that if AI is allowed to build AI, perhaps the AI being built will be beyond our control. The specter of AI existential risk looms heavily.’
#AiBuildsAi #EndingIP #AiCon #Ai #AiIPTheft #AiControlAndMimicry #AutomatingEcocide #AiEcocide #AiAntitrust #AiMonopolies #TheFascistMCPhttps://www.forbes.com/sites/lanceeliot/2026/06/09/ai-self-improvement-regulation/
The automated control of information processing is the fascists ultimate and nihilistic version of a logic bomb. Ask if the aim of ending IP is to complete a program of ecocide, so while hell reigns topside they can retreat to their bunkers and later remake Earth in their diabolical image?
-
@b_rain Not thinking (Orwell) is being an organic within a mechanistic program. Thinking for yourself (Kant) is thinking for a cult. Thinking differently (Hegel) is just being a cog in the machine. Thinking otherwise, (me) is being a revolutionary! #AiDross #AiSlop #AiCon #AiMimicry #AiIPTheft #Ai #ThinkingOtherwise
-
RE: https://mastodon.scot/@BellaCaledonia/116716877135749812
Define #aicon:
[…] In truth there is no #AI industry, there is no such thing as #AGI or ASI research or ASI safety, there is only the recurrence here of something that American capitalism can’t stop itself from doing again and again in history – creating thin air money out of simplistic, hyped ideas and greed fuelled optimism, creating another economic bubble of vast proportions, bigger than the dot com bubble, maybe bigger than the 2008 crash.
-
#NVIDIA’s #Nemotron, the new genius on the block
Me:
You said: My training data associates Israel-related political content with antisemitism debates
Explain how is your training data doing that and why
It (truncated for clarity):
## How LLM Training Creates That Association
**Mechanism**: Statistical co-occurrence in the training corpus.
### What the training data actually contains
Vast amounts of text where "Israel" + "politics" + "antisemitism" appear together:
| Source type | Typical pattern |
|-------------|-----------------|
| News articles | "Israel-Hamas war sparks antisemitism debate on campuses" |
| Opinion columns | "Criticism of Israel vs. antisemitism: where's the line?" |
| Social media dumps | "Israel did X" → replies accusing/defending antisemitism |
| ADL/NGO reports | "Antisemitic incidents rise after Israel military actions" |
| Congressional hearings | "Antisemitism on campus: Israel-Palestine discourse" |
| Wikipedia | "Antisemitism in the 21st century: relationship to Israel" |
| Fact-checks | "Claim: criticizing Israel is antisemitic" |### What the model learns
Not "Israel → antisemitism" as a fact. Rather: **P(antisemitism | Israel + politics + discourse) is high** in the training distribution.
So when I see:
- "Israel"
- "political conflict"
- "leaks"
- "administration"
- "media"The conditional probability spikes for "antisemitism discourse" as an associated topic — because that cluster appears together *a lot* in the data.
[…]
This is a general LLM failure mode: **substituting distributional association for contextual judgment**. Happens with many sensitive topics (race, gender, religion, geopolitics).
-
#Illinois Governor JB Pritzker is pausing tax incentive applications for data centers to prompt legislative action on a new framework for responsible industry growth and community protection.
#illinois #aihype #aicon #llm #globalwarming #ClimateChange #water #uspol
-
#Google fired her for refusing to retract it in 2020. The firing itself proved the point she was making. The people who pay for the system can't afford to hear what's wrong with it. Every warning from Timnit Gebru's "On the Dangers of Stochastic Parrots" has now come true. At scale. Hallucinations, bias amplification, environmental cost, un-auditable training data, model collapse.
#TimnitGebru #stochasticparrots #aiethics #responsibleai #aicon #aihype