Peter Forbes
petersforbes.bsky.social
Peter Forbes
@petersforbes.bsky.social
Live NW of Chicago | Lifelong midwesterner | proud UW Badger
Work: Responsible AI in finance/banking
Posts: AI governance, applied AI/ML, tech, policy, weather, Chicago, assorted stuff and things
Reposted by Peter Forbes
Does anyone want a survivorship bias shortbread
November 29, 2025 at 4:50 AM
Reposted by Peter Forbes
Winter Storm is on for today in Chicago. 5-9 inches predicted. Drive carefully!
November 29, 2025 at 1:56 PM
Reposted by Peter Forbes
Here's a map of the latest NWS forecast showing a widespread, hefty blanket of snow headed for the Midwest this weekend.
November 28, 2025 at 5:06 PM
Reposted by Peter Forbes
November 27, 2025 at 1:19 PM
Reposted by Peter Forbes
Snowy owl at undisclosed location
November 27, 2025 at 2:01 PM
Reposted by Peter Forbes
This explains a lot
This is an insightful but deeply upsetting article about why everyone in the US feels poor, and why the current political situation emerges as a direct result.

www.yesigiveafig.com/p/part-1-my-...
Part 1: My Life Is a Lie
How a Broken Benchmark Quietly Broke America
www.yesigiveafig.com
November 26, 2025 at 11:20 AM
Reposted by Peter Forbes
"LLMs are simply tools that emulate the communicative function of language, not the separate and distinct cognitive process of thinking and reasoning, no matter how many data centers we build."
Is language the same as intelligence? The AI industry desperately needs it to be
The AI boom is based on a fundamental mistake.
www.theverge.com
November 25, 2025 at 2:04 PM
Reposted by Peter Forbes
Take a minute.
this is mesmerizing
November 25, 2025 at 3:02 PM
My father's home town!
November 25, 2025 at 1:49 PM
As an exercise, imagine a day where the NYT reports on the matter at hand in EXACTLY THE CORRECT WAY.

And the underlying problem will still be the same, unmoved by the perfection of the fit between narrative and prior. What is the next step?

Can we just skip to that step? Let's do that.
November 24, 2025 at 10:17 PM
Reposted by Peter Forbes
Some outlets are reporting that DOGE "doesn't exist" anymore. But the fine print on that overturns the whole story: DOGE is just dropping its tarnished brand, not the chainsaw. 🧵
November 24, 2025 at 6:10 AM
Reposted by Peter Forbes
Watching the potential for SNOW this weekend! There will be changes in timing and placement as we fine-tune the forecast through the week, but for those of you traveling... know there is a chance for travel impacts in the region during the weekend.
November 23, 2025 at 10:19 PM
Reposted by Peter Forbes
a few of my favorite snowy owl shots from yesterday in Chicago #birds 🪶
November 23, 2025 at 3:24 PM
Reposted by Peter Forbes
Look at this, a thing of absolute beauty
November 23, 2025 at 4:43 AM
Reposted by Peter Forbes
Happy 13th anniversary to perhaps the funniest NFL play of all time: The Buttfumble
November 22, 2025 at 9:36 PM
Reposted by Peter Forbes
pretty sunset alert
November 23, 2025 at 12:02 AM
As @matt-levine.bsky.social would argue, the optimal level of crime is not 0%, but 10% seems maybe a little high at this scale.
"Meta internally projected late last year that it would earn about 10 per cent of its overall annual revenue — or $16bn — from running advertising for scams and banned goods, internal company documents show" - Reuters www.ft.com/content/4560...
We have to be able to hold tech platforms accountable for fraud
Algorithms ensure that people who click on scams are likely to see more of them
www.ft.com
November 22, 2025 at 3:17 PM
Reposted by Peter Forbes
Last night on LaSalle St.
November 21, 2025 at 8:36 PM
Reposted by Peter Forbes
#PPOD: On 17 March 2023, one of the Copernicus Sentinel-2 satellites captured this rare cloud-free image of Deception Island, an active volcanic caldera located in the South Shetland Islands of Antarctica. Credit: European Union, Copernicus Sentinel-2 imagery 🧪
November 21, 2025 at 4:03 PM
Reposted by Peter Forbes
Absolutely incredible time lapse of yesterday's pyroclastic flows from #Indonesia and the #semuru #volcano @rpappalardo.bsky.social

www.youtube.com/watch?v=Aj0W...
GIANT Pyroclastic Flow from Semeru Volcano – November 19, 2025 (Level IV Alert)
YouTube video by afarTV
www.youtube.com
November 21, 2025 at 6:55 AM
Reposted by Peter Forbes
omfg
Chicago: Did you know that police departments in the Chicago area coordinate on Slack? Now that I know it's accessible to public records requests, request away:

Their server is called CHI-INTEL and requests should be sent to the department that runs it, the Cook County Sheriff.

Send away!
It *is* possible to get cross-jurisdictional and multi-police department Slack chats out of departments via public records requests. Here's a glimpse at the Chicago-area police Slack chat with people responding with the 👀 emoji to the news that there's been an ICE-involved shooting.
November 20, 2025 at 8:28 PM
Certified real one here. Bravo, sir.
Thank you, @marisakabas.bsky.social. Tom Homan is a bitch!!
November 20, 2025 at 5:08 PM
Reposted by Peter Forbes
Nano Banana Pro, released this morning, is clearly the best image generation model. Superb instruction following, plus it can generate full infographics (with correct spelling and properly rendered text!) from a short prompt based on running extra searches simonwillison.net/2025/Nov/20/...
Nano Banana Pro aka gemini-3-pro-image-preview is the best available image generation model
Hot on the heels of Tuesday’s Gemini 3 Pro release, today it’s Nano Banana Pro, also known as Gemini 3 Pro Image. I’ve had a few days of preview access …
simonwillison.net
November 20, 2025 at 4:34 PM
Reposted by Peter Forbes
The site is indeed pure poison, but maybe this is the original article? arxiv.org/pdf/2511.15304
November 20, 2025 at 11:25 AM
Reposted by Peter Forbes
❗ Researchers say they’ve found a universal “jailbreak” for top AI models, through poetry. A new study shows that rephrasing harmful prompts as short poems can bypass safety filters across all major models, raising questions over AI Act compliance.
www.mlex.com/mlex/article...
AI models’ safety features can be circumvented with poetry, research finds | MLex | Specialist news and analysis on legal risk and regulation
A study found that poetic prompts can bypass safety features in leading AI models from OpenAI, Anthropic, Google and others, triggering instructions for building chemical weapons and malware. The rese...
www.mlex.com
November 20, 2025 at 10:54 AM