lol: "As the leader of an AI company which stands to benefit enormously if I convince enough investors that AGI is inevitable, it’s clear to me that AGI is inevitable. But developing superintelligence safely is a […] https://directing.attention.to/p/to-make-ai-safe-we-must-develop-it

"Employees are using AI tools to create low-effort, passable looking work that ends up creating more work for their coworkers: "We define workslop as AI generated work content that masquerades as good work, but […] https://hbr.org/2025/09/ai-generated-workslop-is-destroying-productivity

The winning formula for agrivoltiacs -- very clever. East/west aligned, vertically-mounted solar panels do not impede growing; they provide shelter from wind for the plants; and they provide power when it's […] https://techxplore.com/news/2025-09-harvest-vertical-solar-panels-crops.html

This is actually impressive results from using LLMs to perform security scans on an existing codebase. Daniel Stenberg of curl has given the results of this work a thumbs-up: https://mastodon.social/@bagder/ […] https://joshua.hu/llm-engineer-review-sast-security-ai-tools-pentesters

This has some good points: "Look, it’s starting to be pretty damn obvious that “Free Software” and """Open-Source""" are no longer the kinda hippie shit we thought them to be back when they’d give you Linux distros CDs with magazines […] https://aria.dog/barks/forklift-certified-license/

This is 100% spot on, regarding the never ending series of exploits of failures of npm's security model: "This could be the moment where npm comes to terms with its broken design, and with a well-funded […] https://drewdevault.com/2025/09/17/2025-09-17-An-impossible-future-for-JS.html

"0x.Tools: X-Ray vision for Linux systems". Linux Performance Analysis with Modern eBPF and DuckDB; dig into the captured DuckDB files using "xtop": "xtop is like the Linux top tool, but extended with […] https://tanelpoder.com/posts/xcapture-v3-alpha-ebpf-performance-analysis-with-duckdb/

This is pretty messy. UK companies have taken to outsourcing core IT and infosec to low-cost service providers, then inevitably get hacked -- then make […] https://doublepulsar.com/the-elephant-in-the-biz-outsourcing-of-critical-it-and-cybersecurity-functions-risks-uk-economic-96205e0585bf

Turns out disposable vapes contain a quite capable ARM microcontroller! "So here are the specs of a microcontroller so bad, it’s basically disposable: - 24MHz Coretex M0+; - 24KiB of Flash Storage; - 3KiB of Static RAM; - a few […] https://bogdanthegeek.github.io/blog/projects/vapeserver/

"Reproducibility is a bedrock of scientific progress. However, it’s remarkably difficult to get reproducible results out of large language models. For example, you might observe that asking ChatGPT the same […] https://thinkingmachines.ai/blog/defeating-nondeterminism-in-llm-inference/

Oh great. Russian psyops are now disrupting the fight against malaria: "The move is “a real blow” to hopes for gene drives, says Fredros Okumu, a vector biologist at the […] https://www.science.org/content/article/after-humiliating-raid-burkina-faso-halts-gene-drive-project-fight-malaria

WTF! some TLDs allow anyone to buy the domain _BEFORE_ they expire; e.g. ".pe" allowed a squatter to steal a domain 12 days prior to its expiration. How does this make sense? https://www.namecheap.com/support/knowledgebase/article.aspx/9916/2207/tlds-grace-periods/

LOL. Republican political email campaigns (like WinRed) keep getting marked as spam, because they're using shitty lists: " Tossavainen told KrebsOnSecurity that WinRed’s emails hit its spamtraps in […] https://krebsonsecurity.com/2025/09/gop-cries-censorship-over-spam-filters-that-work/

One dev crunched the numbers on AI coding -- and found absolutely 0 noticeable impact: "I discovered that the data isn’t statistically significant at any meaningful level. That I would need to record new […] https://mikelovesrobots.substack.com/p/wheres-the-shovelware-why-ai-coding

"An investigation by Canada’s National Observer has found that Google’s net-zero pledge has quietly been scrubbed, demoted from having its own section on the site to an […] https://removepaywalls.com/https://www.nationalobserver.com/2025/09/04/investigations/google-net-zero-sustainability

OVHCloud are (rightfully) making plentiful hay from Microsoft's admission that data sovereignty is a joke under US law: ""[ […] https://www.theregister.com/2025/08/27/ovhcloud_interview/?ck_subscriber_id=512829374#438:%20Amazon%20Q%20Rules%20Except%20It%20Doesn't%20At%20All%20-%2018837614

Demand-response actually working in the field! "Artificial intelligence (AI) is fueling exponential electricity demand growth, threatening grid reliability, raising prices for communities paying for new energy infrastructure, and stunting AI innovation […] https://arxiv.org/pdf/2507.00909

Turns out there is an extensive hacking scene turning cars like the Toyota Prius into a full EV with homebrew hardware/firmware (via ITC Slack) https://openinverter.org/forum/viewtopic.php?t=2516

"I've got experience working on censorship circumvention for a major VPN provider" -- good HN comment on this ever-more-relevant topic. Mullvad gets a thumbs up https://news.ycombinator.com/item?id=45055604

This, 100000%: "The “nonprofit” company OpenAI was launched under the cynical message of building a “safe” artificial intelligence that would “benefit” humanity. The company adopted a bunch of science fiction talk popular amongst the […] https://www.argmin.net/p/the-banal-evil-of-ai-safety

Wow OpenAI are really fucking up here. After the truly awful read of the Adam Raine suicide case in the NYT, https://www.nytimes.com/2025/08/26/technology/ […] https://arstechnica.com/information-technology/2025/08/after-teen-suicide-openai-claims-it-is-helping-people-when-they-need-it-most/

Another bizarre behaviour of LLM safety features implemented with logits during post-training: ""Our research introduces a critical concept: the refusal-affirmation logit gap," researchers Tung-Ling "Tony" Li and Hongliang […] https://www.theregister.com/2025/08/26/breaking_llms_for_fun/

For an AWS old-timer user like myself, this list is chock full of "I didn't know that" https://www.lastweekinaws.com/blog/aws-in-2025-the-stuff-you-think-you-know-thats-now-wrong/

Tony Finch: “i accidentally the whole history of email in the 1970s" -- this is great https://lobste.rs/s/gvtlpo/email_is_easy_email_address_quiz#c_vkssdf

This seems spot on: "Using any sort of statistical summary of the data, rather than the aggregated energy and climate impact across the whole system, will always give a misleading view. They […] https://ketanjoshi.co/2025/08/23/big-techs-selective-disclosure-masks-ais-real-climate-impact/

Terrible name, but a serious issue all the same; "Agentic" AI browsers are happily vulnerable to scams and phishing -- " All we did was fake a simple email from a fresh new […] https://guard.io/labs/scamlexity-we-put-agentic-ai-browsers-to-the-test-they-clicked-they-paid-they-failed

A good ol' exfiltration-via-DNS attack. Some day the LLM community will stop reinventing all the classic exploits, I have to assume -- but today is not that day. (Step one in that process would be to […] https://embracethered.com/blog/posts/2025/claude-code-exfiltration-via-dns-requests/

From Mythic Beasts, the UK internet provider: "@cstross to put some numbers on it, one of our hosting VMs has ~1200 mailboxes using 1.5TB of SSD. Accounting for the CPU + RAM to allow the mail to be usable and searchable, you […] https://social.mythic-beasts.com/@beasts/115017750430627832

"Is Chain-of-Thought Reasoning of LLMs a Mirage?": "Chain-of-Thought (CoT) prompting has been shown to improve Large Language Model (LLM) performance on various tasks. With this approach, LLMs appear to produce human-like reasoning steps before providing […] https://arxiv.org/pdf/2508.01191

"The three smart-home hacks are part of a series of 14 indirect prompt-injection attacks against Gemini across web and mobile that the researchers dubbed Invitation Is All You Need. (The 2017 research that led to the recent generative AI […] https://archive.ph/SExCe#selection-1813.0-1829.282

A clever exploit caused by OVPay resolving debits using a nightly batch process: "The exploit is simple. The OVpay processes travel expenses during the overnight hours. Passengers can […] https://nltimes.nl/2023/04/15/scammers-find-quirky-exploit-ovpay-system-free-public-transport-rides

"You probably have not heard Luke Farritor’s name before. He is one of Elon Musk’s 23-year-old DOGE bros who helped dismantle key parts of the federal government, including USAID. The […] https://www.thebulwark.com/p/the-boy-genius-who-killed-14-million-luke-farritor-doge-elon-musk-trump

ETH Zurich are releasing a fully-open AI-Act-compliant large language model: "The model will be fully open: source code and weights will be publicly available, and the training data […] https://ethz.ch/en/news-and-events/eth-news/news/2025/07/a-language-model-built-for-the-public-good.html

Good AI philosophical thoughts via Today In Tabs: "The essential problem is this: generative language software is very good at producing long and contextually informed strings of language, and humanity has never before […] https://www.todayintabs.com/p/we-need-to-talk-about-sloppers-b732

A notable bug from the 2011 Christmas Hurdle at Leopardstown Racecourse: "Even as Voler La Vedette approached the line, the Betfair online market was displaying extremely favorable odds for the horse that was almost certain […] https://kucharski.substack.com/p/a-bit-of-a-christmas-mystery

Quite a clever attack on DMARC; by persuading Google to create a message body that contains the desired phish attack text, then using its legit signing infrastructure to sign the message, an attacker […] https://easydmarc.com/blog/google-spoofed-via-dkim-replay-attack-a-technical-breakdown/

A new interbank instant-payment protocol, to compete with Mastercard/Visa's current monopoly, being rolled out by a group of EU banks (via Abban) https://wero-wallet.eu/

"There's a trend of reassuring people about this by asking spirits like Asmodeus the Prince of Lies if they are being truthful. This feels naive at best and actively malicious at worst." Genius -- there is indeed a lot of commonality […] https://merveilles.town/@d6/114905935294392115

Interesting artifact of training your speech-to-text tool on Youtube videos: 'Complete silence is always hallucinated as "ترجمة نانسي قنقر" in Arabic which translates as "Translation by Nancy Qunqar"' https://github.com/openai/whisper/discussions/2608

A good blog post on Psychology Today regarding the the deepening problem of LLM chatbots creating cases of psychosis in the population. Part of the problem with the LLMs is a failure by […] https://www.psychologytoday.com/us/blog/urban-survival/202507/the-emerging-problem-of-ai-psychosis

Turns out "flooding the zone with shit" isn't just a Trumpian tactic, it's used by fossil fuel anti-climate groups and companies too: "An independent analysis of 45 right-wing groups advocating against trans rights found […] https://heated.world/p/fossil-fuel-billionaires-are-bankrolling

This is a cool phone feature. “AEA demonstrates that globally distributed smartphones can be used to detect earthquakes and issue warnings at scale with an effectiveness comparable to […] https://gizmodo.com/android-phones-can-detect-earthquakes-before-the-ground-starts-shaking-2000629636

We've known this for years, but it's significant to see Microsoft admit this under oath in a European court. When the US Government issues an NSL, Microsoft cannot say no: "Microsoft France's legal director […] https://ppc.land/microsoft-cant-protect-french-data-from-us-government-access/

Excellent description of the layers of tuning available for LLMs, and the risks involved, as demonstrated by Grok's recent "MechaHitler" incident: "LLMs become “woke” because they are trained to be pro- […] https://prospect.org/power/2025-07-17-how-did-elon-musk-turn-grok-into-mechahitler/

Delta's going to start charging based on "AI", through "a partnership with Fetcherr, a six-year-old Israeli company that also counts Azul, […] https://fortune.com/2025/07/16/delta-moves-toward-eliminating-set-prices-in-favor-of-ai-that-determines-how-much-you-personally-will-pay-for-a-ticket/

Cory Doctorow on how Google are desperate to maintain a facade of being a "growth" company: "Investors have metabolized the story that AI will be a gigantic growth area, and so all the tech giants are in a battle to prove to investors […] https://pluralistic.net/2025/07/15/inhuman-gigapede/

Measurements of the effectiveness of the "Pravda" disinformation network: "Even with [knowledge of LLM Grooming], ChatGPT nevertheless often repeats propaganda from Pravda. Model o3, OpenAI’s allegedly state […] https://americansunlight.substack.com/p/bad-actors-are-grooming-llms-to-produce

Wow, this is really shocking stuff -- and I have to say, not surprising to me at all: "Over a […] https://lyons-den.com/whitepapers/aws-lambda-silent-crash.pdf?ck_subscriber_id=512829374#431:%20AWS%20Sovereign%20Cloud,%20Now%20With%20Slightly%20More%20Pretend%20Sovereignty%20-%2018285791

Sites and services which have closed up in the UK due to the risks imposed by the introduction of the Online Safety Act 2023. Meanwhile, the Act has explicit exemptions for "news publishers" and the comments below their articles -- ie. the […] https://onlinesafetyact.co.uk/in_memoriam/

Groupthink underpinned the flawed thinking behind the UK’s pandemic response, a succession of witnesses at the heart of government told the Covid-19 public […] https://www.theguardian.com/uk-news/article/2024/jul/18/covid-inquiry-hallett-prescribes-red-teams-as-antidote-to-flawed-thinking