bolha.us

73 posts67 participants2 posts today

PrivacyDigest<a href="https://mas.to/tags/Facebook" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Facebook</a> Pushes Its <a href="https://mas.to/tags/Llama4" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Llama4</a> <a href="https://mas.to/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> Model to the Right, Wants to Present “Both Sides” <a href="https://mas.to/tags/Bias" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Bias</a> in artificial intelligence systems, or the fact that large language models, facial recognition, and <a href="https://mas.to/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> image generators can only remix and regurgitate the information in data those technologies are trained on, is a wellestablished fact that researchers and academics have been warning about since their inception. <a href="https://mas.to/tags/maga" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#maga</a> <a href="https://mas.to/tags/right" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#right</a> <a href="https://mas.to/tags/bothsides" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#bothsides</a> <a href="https://mas.to/tags/llm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#llm</a> <a href="https://mas.to/tags/facialrecognition" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#facialrecognition</a> <a href="https://www.404media.co/facebook-pushes-its-llama-4-ai-model-to-the-right-wants-to-present-both-sides/" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://www.404media.co/facebook-pushes-its-llama-4-ai-model-to-the-right-wants-to-present-both-sides/</a>

@990000@mstdn.socialHaha nice <a href="https://mstdn.social/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://mstdn.social/tags/CyberPunk" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#CyberPunk</a> <a href="https://mstdn.social/tags/Malware" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Malware</a> <a href="https://bsky.app/profile/thomasfuchs.at/post/3lmnqvt35bk2z" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://bsky.app/profile/thomasfuchs.at/post/3lmnqvt35bk2z</a>

Paco Hope #resistMore lessons learned using an <a href="https://infosec.exchange/tags/ai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ai</a> <a href="https://infosec.exchange/tags/chatbot" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#chatbot</a> to help write code. (See <a href="https://infosec.exchange/@paco/114316516639964570" rel="nofollow noopener noreferrer" target="_blank">my original post</a> for some of the other observations).First, watch out for vibe coding. I start off asking it to do things I know and things I could double check if I wanted to. And it's really easy to start asking it to do things that I don't know how to do, and that I can't check.Second, save checkpoints. I'm inviting the <a href="https://infosec.exchange/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> to modify the code files. Sometimes it modifies unrelated stuff that used to work and makes it not work, or makes it different. I ask it to calculate alt text, and it changes how the histogram is plotted. As I iteratively get things working, I do a <code>git add</code> to capture a working version. A couple times I let it make a series of changes and by the time I got to the last change, it had clobbered some intermediate states that were working, and I didn't have a copy of the code that was there previously. At each incremental step, VS Codium can use git information to highlight which lines have changed due to the most recent modifications.Last, it goes without saying, I must go in behind it and clean things up. A funny thing it tends to do is get really attached to adjectives. I wanted my trend line to be yellow, and in order to be clear, I prompted with "add a 15-minute moving average trend line that is bright yellow". Everywhere it mentioned this trend line (in comments, in the alt text) it always said "a bright yellow trend line" instead of just "a yellow trend line". When I asked it to make the histogram bins "exactly one minute" (again, trying to be unambiguous), it then labels everything like the X axis label and the alt text with the phrase "exactly 1 minute" instead of just "1 minute".

IT NewsVibe Check: False Packages a New LLM Security Risk? - Lots of people swear by large-language model (LLM) AIs for writing code. Lots of p... - <a href="https://hackaday.com/2025/04/12/vibe-check-false-packages-a-new-llm-security-risk/" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://hackaday.com/2025/04/12/vibe-check-false-packages-a-new-llm-security-risk/</a> <a href="https://schleuss.online/tags/artificialintelligence" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#artificialintelligence</a> <a href="https://schleuss.online/tags/securityhacks" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#securityhacks</a> <a href="https://schleuss.online/tags/softwarehacks" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#softwarehacks</a> <a href="https://schleuss.online/tags/codesecurity" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#codesecurity</a> <a href="https://schleuss.online/tags/llm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#llm</a> <a href="https://schleuss.online/tags/ai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ai</a>

⚛️Revertron<a href="https://zhub.link/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://zhub.link/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a>

RenézuCode@AMD Releases <a href="https://chaos.social/tags/ROCm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ROCm</a> 6.4: More Powerful, More Modular, but Still No RDNA4 Support <a href="https://medium.com/p/amd-releases-rocm-6-4-more-powerful-more-modular-but-still-no-rdna4-support-ff3030d3da3a?source=social.tw" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://medium.com/p/amd-releases-rocm-6-4-more-powerful-more-modular-but-still-no-rdna4-support-ff3030d3da3a?source=social.tw</a> <a href="https://chaos.social/tags/opensource" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#opensource</a> <a href="https://chaos.social/tags/llm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#llm</a> <a href="https://chaos.social/tags/ai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ai</a> <a href="https://chaos.social/tags/hpc" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#hpc</a>

Toni Aittoniemi<a href="https://mastodon.world/@FAIR" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@FAIR</a> Oh yes. They started blocking pro-Palestine and anti-Russia messaging on Ukraine for me a long time ago. This is of course systematic abuse of reporting tools by automated means or by troll farms in the case of Kremlin.Facebook refuses to hire enough people to counter the problem, which is just getting exponentially worse by cheap LLM’s able to do the job which you formerly needed a Russian spy for. <a href="https://mastodon.green/tags/ai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ai</a> <a href="https://mastodon.green/tags/disinformation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#disinformation</a> <a href="https://mastodon.green/tags/llm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#llm</a> <a href="https://mastodon.green/tags/botnet" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#botnet</a> <a href="https://mastodon.green/tags/socialmedia" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#socialmedia</a> <a href="https://mastodon.green/tags/facebook" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#facebook</a>

randomprecisionA2A protocol <a href="https://lemmy.world/post/28112560" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://lemmy.world/post/28112560</a>

TheZeldaZone👑🏳️‍⚧️🎮🎀I do not consent to the <a href="https://mastodon.social/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> scraping this data.

Paco Hope #resistOk. Another funny <a href="https://infosec.exchange/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> output while asking the <a href="https://infosec.exchange/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> to suggest changes to my code. I asked it to recommend a trend line for my histogram graph.I want my histogram to have a dark background, so I found <a href="https://github.com/akasharidas/dark-matplotlib-styles" rel="nofollow noopener noreferrer" target="_blank">this blog post</a> full of dark matplotlib themes and I picked one I liked. My graphing code has both <code>import mplcyberpunk</code> to bring in the style and later <code>plt.style.use('cyberpunk')</code> when I do my histogram plotting.In recommending a trend line to me, Amazon Q Developer says:<blockquote>For your use case, I recommend Option 1 (Moving Average) because:<ol><li>It's simple to implement without additional dependencies</li><li>It works well with time series data showing activity over time</li><li>It's easy to adjust the window size to control smoothness</li><li>It fits with the cyberpunk aesthetic by adding a smooth line over the histogram</li><li>It's computationally efficient</li></ol></blockquote>Who knew that having a smooth line over a histogram would fit a "cyberpunk aesthetic"?

Blake PattersonThinking Different, Thinking Slowly: LLMs on a PowerPC Mac [The Resistor Network]<a href="http://www.theresistornetwork.com/2025/03/thinking-different-thinking-slowly-llms.html" rel="nofollow noopener noreferrer" translate="no" target="_blank">http://www.theresistornetwork.com/2025/03/thinking-different-thinking-slowly-llms.html</a><a href="https://oldbytes.space/tags/Mac" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Mac</a> <a href="https://oldbytes.space/tags/vintageMac" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#vintageMac</a> <a href="https://oldbytes.space/tags/PowerBook" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#PowerBook</a> <a href="https://oldbytes.space/tags/PowerPC" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#PowerPC</a> <a href="https://oldbytes.space/tags/PPC" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#PPC</a> <a href="https://oldbytes.space/tags/AltiVec" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AltiVec</a> <a href="https://oldbytes.space/tags/G4" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#G4</a> <a href="https://oldbytes.space/tags/G5" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#G5</a> <a href="https://oldbytes.space/tags/vintagecomputing" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#vintagecomputing</a> <a href="https://oldbytes.space/tags/retrocomputing" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#retrocomputing</a> <a href="https://oldbytes.space/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://oldbytes.space/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://oldbytes.space/tags/machinelearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#machinelearning</a> <a href="https://oldbytes.space/tags/coding" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#coding</a> <a href="https://oldbytes.space/tags/Apple" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Apple</a> <a href="https://oldbytes.space/tags/tech" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#tech</a> <a href="https://oldbytes.space/tags/blog" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#blog</a> <a href="https://oldbytes.space/tags/programming" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#programming</a>

PositThe latest posit::glimpse() newsletter is out! Read about new <a href="https://fosstodon.org/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> tools for <a href="https://fosstodon.org/tags/RStats" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#RStats</a> and <a href="https://fosstodon.org/tags/Python" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Python</a>, improved Python/R interoperability, community events, learning resources, and more.<a href="https://posit.co/blog/posit-glimpse-newsletter-april-2025/" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://posit.co/blog/posit-glimpse-newsletter-april-2025/</a>

Alessio Pomaro🧠 Cognition lancia la nuova versione di <a href="https://mastodon.uno/tags/Devin" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Devin</a>, con delle novità interessanti per i developer. 👉 Le nuove funzionalità: <a href="https://www.linkedin.com/posts/alessiopomaro_devin-ai-genai-activity-7316417567923875841-10mr" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://www.linkedin.com/posts/alessiopomaro_devin-ai-genai-activity-7316417567923875841-10mr</a>___ ✉️ 𝗦𝗲 𝘃𝘂𝗼𝗶 𝗿𝗶𝗺𝗮𝗻𝗲𝗿𝗲 𝗮𝗴𝗴𝗶𝗼𝗿𝗻𝗮𝘁𝗼/𝗮 𝘀𝘂 𝗾𝘂𝗲𝘀𝘁𝗲 𝘁𝗲𝗺𝗮𝘁𝗶𝗰𝗵𝗲, 𝗶𝘀𝗰𝗿𝗶𝘃𝗶𝘁𝗶 𝗮𝗹𝗹𝗮 𝗺𝗶𝗮 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿: <a href="https://bit.ly/newsletter-alessiopomaro" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://bit.ly/newsletter-alessiopomaro</a> <a href="https://mastodon.uno/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://mastodon.uno/tags/GenAI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#GenAI</a> <a href="https://mastodon.uno/tags/GenerativeAI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#GenerativeAI</a> <a href="https://mastodon.uno/tags/IntelligenzaArtificiale" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#IntelligenzaArtificiale</a> <a href="https://mastodon.uno/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://mastodon.uno/tags/coding" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#coding</a>

LeoGrok 3 API finally dropped. Time to add it to my utilities and try it out.The pricing of the large model is the same as Claude. Let's hope it can back it up by providing a similar quality.<a href="https://masto.gkbrk.com/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://masto.gkbrk.com/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://masto.gkbrk.com/tags/Grok" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Grok</a>

Aho<a href="https://mastodon.social/@arstechnica" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@arstechnica</a> the ting I wish was they took away the laziness in the later <a href="https://ioc.exchange/tags/llm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#llm</a> as if I ask to see a code example for something, why would I want some list of steps to do which mostly telling you to search the internet yourself for the answers.

John Gordon 📷One of the parents on my daughter's softball team worked out a <a href="https://mastodon.social/tags/ChatGPT" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ChatGPT</a> prompt that produced fun images of an action figure version of a player in packaging. I am now asking it to generate me 3D printable STL files from this image I got from the prompt.<a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://mastodon.social/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> <a href="https://mastodon.social/tags/3dPrinting" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#3dPrinting</a>

Paco Hope #resistLast thing that I thought was silly was the <a href="https://infosec.exchange/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a>'s documentation on alt text. I'm a big fan of alt text and I think it is important. That's why I'm putting it in there.And boy, howdy, did the LLM think it was special. When I asked it to add documentation on the new alt text feature, it spontaneously adds this silly line about "Remember this is a work in progress and we will keep making it more accessible!" It added several lines in several places that just emphasized how happy it was to be writing alt text. And this is just that sickening <a href="https://infosec.exchange/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> cheerfulness. Like when you prompt it saying "that code doesn't work" and it shamelessly replies "You're right! I should have done XYZ instead of the stupid shit that I originally said would work."Really, alt text should just be the norm. Things that lack accessibility should stand out as crude, unpolished, and not worth using. It shouldn't be that software that has accessibility is held out as especially good. Accessibility should be table stakes.10/n

Paco Hope #resistI did have it do something that I didn't know exactly how to do, which was get some statistics out of my histogram graph. It picked a stupid date variable that was a string (like <code>"Sun 8 Apr 2025"</code>) and was trying to use that in a place that expected an integer.I was able to paste the error and say<blockquote>That doesn't work. I get this error.</blockquote>And like any cheerful intern, it blithely admits that it totally fucked that up and it tries a fix. But the first fix it tried it assumed some kind of divide-by-zero error, which wasn't right. On the second attempt it actually produced code that worked.Then there's this peculiar line of Python. This is how it avoided the divide by zero problem:<pre><code>avg_posts_per_minute = len(event_timestamps) / max(event_duration, 0.001) # Avoid division by zero </code></pre>While it is true that <code>max(event_duration, 0.001)</code> will avoid divide-by-zero if <code>event_duration</code> is 0 or negative, dividing by <code>0.001</code> is multiplying by 1000! So this is not really how I want to handle the situation where <code>event_duration</code> is zero. That's a situation to handle, and not with silly math that makes it produce a crazy value instead of exploding. That's kinda worse, in some ways.And this leads me to a meta-gripe about all this <a href="https://infosec.exchange/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a> generated code: it tries to hard to keep going. Let me explain:7/n

Reilly Spitzfaden (they/them)“Most computer vision tech ultimately serves as some form of surveillance, sentencing algorithms discriminating against Black people, and a primary driver of AI generated images, video, and audio is nonconsensual media of women.”<a href="https://www.404media.co/facebook-pushes-its-llama-4-ai-model-to-the-right-wants-to-present-both-sides/" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://www.404media.co/facebook-pushes-its-llama-4-ai-model-to-the-right-wants-to-present-both-sides/</a><a href="https://hachyderm.io/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://hachyderm.io/tags/Surveillance" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Surveillance</a> <a href="https://hachyderm.io/tags/Meta" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Meta</a> <a href="https://hachyderm.io/tags/Racism" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Racism</a> <a href="https://hachyderm.io/tags/LLMs" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLMs</a> <a href="https://hachyderm.io/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a>

nixCraft 🐧Fake job seekers are flooding U.S. companies that are hiring for remote positions, tech CEOs say <a href="https://www.cnbc.com/2025/04/08/fake-job-seekers-use-ai-to-interview-for-remote-jobs-tech-ceos-say.html" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://www.cnbc.com/2025/04/08/fake-job-seekers-use-ai-to-interview-for-remote-jobs-tech-ceos-say.html</a> I didn't know fake candidates were a thing?!<a href="https://mastodon.social/tags/security" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#security</a> <a href="https://mastodon.social/tags/jobs" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#jobs</a> <a href="https://mastodon.social/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> <a href="https://mastodon.social/tags/LLM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLM</a>

Recent searches

Search options

Administered by:

Server stats:

#llm