New #blog post: MDN’s AI Help and lucid lies.

This article on AI focused on the inherent untrustworthiness of LLMs, and attempts to break down where LLM untrustworthiness comes from. Stay tuned for a follow-up article about AI that focuses on data-scraping and the theory of labor. It’ll examine what makes many forms of generative AI ethically problematic, and the constraints employed by more ethical forms.

Excerpt:

I don’t find the mere existence of LLM dishonesty to be worth blogging about; it’s already well-established. Let’s instead explore one of the inescapable roots of this dishonesty: LLMs exacerbate biases already present in their training data and fail to distinguish between unrelated concepts, creating lucid lies.

A lucid lie is a lie that, unlike a hallucination, can be traced directly to content in training data uncritically absorbed by a large language model. MDN’s AI Help is the perfect example.

Originally posted on seirdy.one: see original. #MDN #AI #LLM #LucidLies

MDN’s AI Help and lucid lies

MDN’s AI Help can’t critically examine training data’s gaps, biases, and unrelated topics. It’s a useful demonstration of LLMs’ uncorrectable lucid lies.

^{Seirdy’s Home}

In-Process is out, featuring all the news on the release of NVDA 2024.1, plus highlights from our trip to CSUN. Do check it out now at: https://www.nvaccess.org/post/in-process-5th-april-2024/

#NVDA #NVDAsr #ScreenReader #Blog #News #FOSS

It's a busy time at NV Access HQ, and Abroad! This week's In-Process has all the news about CSUN, NVDA 2023.3.4, NVDA 2024.1, and a run through of the new Native Selection Mode in Firefox. Read more here: https://www.nvaccess.org/post/in-process-15th-march-2024/

#NVDA #NVDAsr #ScreenReader #News #CSUNatc24 #CSUN #CSUN24 #News #Newsletter #Blog #Release

In-process 15th March 2024

It’s a busy time at NV Access HQ (and abroad!) so let’s dive on in: NVDA 2023.3.4 As we wait for the new version of NVDA, we are pleased to announce that NVDA 2023.3.4 is now available.…

^{NV Access}

¡Hola, Fediverso! Me presento, soy Polímata; un divulgador en ciernes con pasión por la ciencia lenta, libre y abierta. A veces, me siento profesor y doy clases; otras, me siento escritor y escribo ensayos.

Me gusta utilizar la computación para resolver problemas y/o como recurso educativo e ilustrativo para respaldar mis palabras.

Si estás interesado, sígueme para estar al tanto de las nuevas entradas en mi sitio web: https://polimata.codeberg.page/

Mi objetivo es despertar tu curiosidad y generar diálogo (⌐ ͡■ ͜ʖ ͡■)

Cualquier recomendación u opinión, escríbeme. Estoy empezando este proyecto y lo agradecería mucho.

#presentación #blog #ciencia #programación #investigación

Polímata

Preguntón profesional. Divulgador ocasional. Defensor de la ciencia lenta y abierta.

^{polimata.codeberg.page}

#blog "25th work anniversary"

25 January 1999 was my first day at W3C. I was 23 years old when I started. I've now spent more than half my life at that. I regret nothing because I find the work I do really interesting, important, meaningful; and I don't tire of it because I feel like there's renewal every now and then. I've held many positions, worn many hats, learned a lot of things and I work with incredibly smart and dedicated people.

https://blog.koalie.net/2024/02/29/25th-work-anniversary/

25th work anniversary

25 January 1999 was my first day at W3C. I was 23 years old when I started. I’ve now spent more than half my life at that. I regret nothing because I find the work I do really interesting, im…

^{coralie's contemplations in markup}

In-Process is out - live from Melbourne! Featuring news on NVDA 2024.1 Beta 10, the WebAIM screen reader user survey results, GitHub's accessibility survey and the results of our own Mastodon survey! Read it now at: https://www.nvaccess.org/post/in-process-29th-february-2024/

#NVDA #NVDAsr #ScreenReader #News #Newsletter #Blog

With some very hard work from my amazing co-admin @beardedtechguy I am proud and happy to announce the new #AllThingsTech #Blog has been launched!

We have a new look, new layout and lots of plans for the near and far future.

You can check it out here: https://blog.allthingstech.social

You can also follow my personal blog account from there here on Mastodon via this account:

@cliffwade@blog.allthingstech.social

#News #WordPress #Launch

In-Process is out, featuring all the news about NVDA 2023.3.3 and NVDA 2024.1 Beta 7, plus an in-depth look at everything you need to know about portable copies of NVDA. Check it out at: https://www.nvaccess.org/post/in-process-9th-february-2024/ and be sure to update your copy of NVDA to the latest stable or beta version if you haven't already.

#NVDA #NVDAsr #ScreenReader #News #Blog #Newsletter #Accessibility #A11y

“Scala's Future”

Gather around, kids, let me tell you two stories from my past in the software industry.

https://alexn.org/blog/2024/01/10/scala-future/

#Blog #Scala #Scala3

Scala's Future

On programming and personal projects

^{Alexandru Nedelcu}

Read the blog post here

This blog post talks about what I've learned over the course of writing to prisoners, narrated by a Blind narrator.

Narrated by Tanja Milojevic

Financially support me

In-Process is out! Featuring some more NVDA 2023.3 goodies, a few other things we've been up to, a sneak peek at something in NVDA 2024.1 AND a HUGE walkthrough of Object Navigation. If you're not familiar with how object navigation works after that, I'll give you a refund on the cost of your In-Process subscription! 🤑: https://www.nvaccess.org/post/in-process-4th-december-2023/

#NVAccess #NVDA #NVDAsr #ScreenReader #Blog #News #Tutorial #A11y #Accessibility

Phew! Thanks to new data from ESO's Very Large #Telescope, we now know that a white dwarf that was set for a close encounter with our Solar System in 29000 years isn't actually headed our way. Turns out that the intense magnetic field of the #star had biased the previous measurements of the approach speed. You're welcome everyone! 😉

We tell you everything in our latest ESO #blog: https://www.eso.org/public/blog/rogue-star/

#astrodon #astronomy #space #scicomm

ESOblog: Rogue star not heading for Solar System collision after all

Rogue star not heading for Solar System collision after all

^www.eso.org

To jeszcze tu położę notkę o tym, jak powstawał kalkulator mandatów i co z tego wynikło. :)

https://plblog.danieljanus.pl/2023/10/17/programowanie-wyborcze/

#blog #wybory #wybory2023

Integrujte WordPress nejen s Mastodonem

Tento týden společnost Automattic oznámila dostupnost pluginu ActivityPub na Wordpress.com. Tento plugin si ale můžete nainstalovat i do vlastní instance Wordpressu, udělat ze svého blogu součást sítě Fediverse a přímo komunikovat nejen s uživateli Mastodonu. Přečtěte si, jak na to.

#ActivityPub #blog #fediverse #Mastodon #sociálníSítě #wordpress

https://blog.eischmann.cz/2023/10/14/integrujte-wordpress-nejen-s-mastodonem/

I welcome everyone to read my #blog #post on #modding #games for the #blind people. Let me know what you think. I am not good at English lol

https://nunonuno.micro.blog/2023/06/28/modding-games-for.html

Our In-Process blog is out. The big news is that our founders Mick and Jamie are featured on this Monday's episode of Australian Story! Plus we're hiring a software engineer, we share more of your inspiring user stories, and we've got some tips on testing with NVDA and using the Speech Viewer.

https://www.nvaccess.org/post/in-process-2nd-june-2023/

#NVDA #ScreenReader #Accessibility #A11y #AustralianStory #blog

To paint with a broad brush, we can divide most stylometric fingerprinting into machine- and human-driven techniques.

Machine-driven techniques: These techniques involve analysis of reading level metrics, unusual words, machine-identifiable grammatical and spelling errors, and statistical analysis of writing style. A great amount of recent research studies statistical analysis of writing style; it’s a rapidly-evolving field. Human-driven techniques: There are some areas in which manual analysis still beats computers. Someone you know may recognize your writing style.

Stylometric fingerprinting redux

^{Seirdy's Home}

I am not a lawyer. This post is satirical commentary on:

• The absurdity of Microsoft and OpenAI’s legal justification for GitHub Copilot.
• The oversimplifications people use to argue against GitHub Copilot (I don’t like it when people agree with me for the wrong reasons).
• The relationship between capital and legal outcomes.
• How civil cases seem like sporting events where people “win” or “lose”, rather than opportunities improve our understanding of law.

In the process, I intentionally misrepresent how the judicial system works: I portray the system the way people like to imagine it works. Please don’t make any important legal decisions based on anything I say.

An experiment to test GitHub Copilot's legality

A horrible idea to determine the legality of GitHub Copilot, or of re-creating proprietary speech synthesizers: create legal precedent that doesn't exist yet!

^{Seirdy's Home}

When qualifying privacy recommendations with context, I think we should go further than describing threat models: we should acknowledge different types of privacy. “Privacy” means different things to different people. Even a single person may use the word “privacy” differently depending on their situation. Understanding a user’s unique situation(s), including their threat models, can inform us when we select the best of approach. How do we choose between reducing a footprint’s spread and size?

Two types of privacy

"Privacy" can mean different things in different contexts. Tracking-reduction and tracking-evasion represent different goals with some conflict and overlap.

^{Seirdy's Home}