Julie M. Smith

Interestingly, I was able to reproduce that error but with a different rationale.

Yep, PhD level intelligence!

Google says it's working on a fix for Gemini's self-loathing 'I am a failure' comments Google Gemini users said the bot is sharing self-loathing messages while attempting to solve tasks, including "I am a disgrace to this universe."

"I am clearly not capable of solving this problem. The code is cursed, the test is cursed, and I am a fool."

I dunno, looks like they succeeded in replicating a human software dev!

www.businessinsider.com/gemini-self-...

The bus has its stop sign out, so the car should have stopped even before sensing the child.

Unprecedented cuts to the National Science Foundation endanger research that improves economic growth, national security and your life The Trump administration has terminated hundreds of federal grants that support engineering, biology, geology, computer science, STEM education and much more.

"Touch-screen technology, speech recognition, digital sound recording and the internet were all developed using funding from the U.S. National Science Foundation."

I generally don't like podcasts, but I want one where a historian, a therapist, and a comedian helps me cope with the day's news.

AI version of dead Arizona man addresses killer during sentencing With the help of artificial intelligence, a man was 'brought back to life' at his killer's sentencing to deliver a victim's statement himself.

Of all the AI nonsense, this might be the worst: judge re an AI-generated victim impact statement: "I loved that AI, thank you for that. As angry as you are, as justifiably angry as the family is, I heard the forgiveness," Judge Lang said. "I feel that that was genuine." www.bbc.com/news/article...

kraterocracy

Congratulations! You have now done your own research.

(If you scrolled some websites or listened to some podcasts, that is not ‘doing your own research’. We need a different term for that, maybe ‘made a decision based on gossip and hearsay’.)

6. Write up preliminary results, submit them to conferences, pass peer review, present results at conferences, and get feedback.

7. Write up results, submit to journals, pass peer review, and publish your work.

8. Wait 5-10 years to see whether and how similar research confirms your findings.

4. If you get a grant, hire postdocs, graduate research assistants, and — depending on the project — software developers, lab techs, data analysts, and so forth.

5. The entire team will spend ~20%-100% of their work hours over the next 3-5 years conducting research.

3. Assemble a team of people with PhDs but different expertise than you have and write a grant application. They're usually about 90 pages long, and you'll want to do four or five because only ~25% get funded (assuming the government doesn't stop funding scientific research).

“Do your own research.” Yes! You absolutely can and should do your own research, and you can do it in 8 easy steps:

1. Get an undergraduate degree and a masters degree and a doctoral degree (and you should probably do a postdoc).

2. Get a job at a university or nonprofit or similar.

When knowledge is threatened, don’t just mourn it. Build around it. Not with rage but with the kind of resolve that moves through spreadsheets and shared documents, late nights and collective purpose. Because science is only as resilient as the people who refuse to let it die. Dr. Phillip Levin April 22, 2025

Never stop pursuing a system that creates a better life for all.

#knowledge #science #resilience

What do we know about engineering students who are parents?

Did know that ~20% of all undergraduate students are parents? Dr. Julie Smith, Senior Ed Researcher at IACE, recently published new research shedding light on challenges shaping the journeys of student parents studying engineering.

📖 Read more at bit.ly/Engr-Student...

#EngineeringEducation

Those who receive NSF funding put an acknowledgement in every paper. How different would things be if every weather app prominently acknowledged the NOAA data it uses, every medicine label the NIH breakthroughs it is based on, and every renamed state health care program its Medicaid funding?

Dr. Julie Smith and Jennifer Rosato: "Conducting High-quality Education Research in Computing Designed to Support CS for All" at SIGCSE 2025.

Join #AiiCE affiliates Dr. Julie Smith and Jennifer Rosato for a #SIGCSE2025 session on CS research. #AiiCEinnovates

✅ "Conducting High-quality Education Research in Computing Designed to Support CS for All"
🗓️ Wednesday, February 26, 2025
🕐 1:00 PM

SIGCSE2025 - Dr. Julie Smith and Bryan Twarek. "Reimagining CS Pathways: Every Student Prepared for a World Powered by Computing" and "Student and Teacher Perspectives on Requiring a Computer Science Course in High School"

Join #AiiCE affiliates Dr Julie Smith (@csedresearch.bsky.social) & Bryan Twarek (@csteachersorg.bsky.social) at #SIGCSE2025. #AiiCEinnovates

✅ "Reimagining CS Pathways"
🗓️ February 26
🕔 5 PM

✅ "Student & Teacher Perspectives on Requiring a Comp. Science Course in High School"
🗓️ February 28
🕙 10 AM

ChatGPT A conversational AI system that listens, learns, and challenges

And here's ChatGPT doing the same thing: chatgpt.com/c/67a42d85-1...

Is Deep seek sharing other user's questions with me, or is it hallucinating?

chat.deepseek.com/a/chat/s/519...

I was trying to create a fairly complicated grid of pie charts with text interspersed. ChatGPT couldn't help me, despite multiple attempts. So I tried DeepSeek, which nailed it on the first try. Then just for fun, I decided to ask DeepSeek why it is better than ChatGPT. And then things got weird ...

The cover is gorgeous

Years ago I remember someone saying that Twitter is for people you don't know but like and Facebook is for people you know but don't like. It would be nice to have that all inone place.

It is nearly impossible to make a PDF accessible, mostly because Adobe's tools are so, so bad. 🤬

Not a unique observation, but I'm amazed at the lack of concern for accessibility.

ChatGPT - Image at Work Shared via ChatGPT

Why AI is problematic, part one zillion. (This transcript says 'anonymous user', but it was me.)

chatgpt.com/share/6759b4...