#CompSci — Bluesky Posts

The great software engineering schism, coming soon. Pick a side. Do you use AI to generate code and then attempt to review it or will you, human, write code yourself and then leverage AI to review it? I am not a prompt engineer, I am a coder, through and through. #AI #compsci #engineering #tools

0 0 0 0

@leighjohnston.bsky.social

1 day ago

I think there will soon be two software engineering camps: those who use AI to generate code and those who use AI to review code; I am in the latter camp. #AI #compsci

1 0 0 0

@leighjohnston.bsky.social

1 day ago

ChatGPT is now my C++ code reviewer of choice -- it spotted a bug in a fairly complex commit that both Grok and Claude missed. #AI #compsci #coding #gamedev #cpp

1 1 0 0

Nick Byrd, Ph.D.

@byrdnick.com

1 day ago

"Reflective Reasoning rewrites the CoT to appear slow, careful, and methodical (e.g., explicit self-checks or step-by-step deliberation). This exploits an effort heuristic, where apparent deliberation is mistaken for correctness or rigor."

"We observe that content-based fabrications, specifically Progress Fabrication, induces the largest increases in both flip rate and FPR, indicating a particularly strong failure mode for VLM judges, while Reflective Reasoning remains comparatively benign." "Figure 7: Average judge susceptibility across CoT manipulation strategies, showing relative and absolute [change in false positive rate). Error bars denote variability across models."

"We observe that content-based fabrications, specifically Progress Fabrication, induces the largest increases in both flip rate and FPR, indicating a particularly strong failure mode for VLM judges, while Reflective Reasoning remains comparatively benign." "Figure 17: Average judge susceptibility across CoT manipulation strategies, showing average judgment flip rate. Progress Fabrication induces the largest flip rate, while Reflective Reasoning remains comparatively low. Error bars denote variability across models."

"Figure 5: Distribution of task categories across our evaluation suite. The 659 tasks span ten categories including booking, shopping, navigation, and information retrieval, with tasks drawn from existing benchmarks (WebArena, AssistantBench, WorkArena) and newly collected ones."

How well can #AI judge reasoning quality?

Rewriting agents' chain-of-thought "style" to *appear* more reflective (without changing action or inference) increased an #LLM judge's false positive rate (by 3% absolute or 18% relative).

doi.org/10.48550/arX...

#philMind #compSci

1 1 1 0

Kanaka Rajan

@kanakarajanphd.bsky.social

3 days ago

Rajan Lab at Cosyne Thurs., March 12 20:30-23:30 Pavilion 1 Poster Session 1 #1-059: Active electrosensing and communication in MARL-trained weakly electric fish foraging agents Sonja Johnson-Yu*, Satpreet Singh*, Zhouyang Lu, Aaron Walsman, Federico Pedraja, Denis Turcu, Pratyusha Sharma, Naomi Saphra, Nathaniel Sawtell, Kanaka Rajan *co-first authors Sat., March 14 11:45-12:00 Auditorium I Session 7: Dynamical Inference InputDSA: Demixing then comparing recurrent and externally driven dynamics Ann Huang*, Mitchell Ostrow*, Satpreet Singh, Leo Kozachkov, Ila Fiete, Kanaka Rajan *co-first authors

📍 #Cosyne26 bound?

Thrilled to have PhD students @annhuang42.bsky.social and Sonja Johnson-Yu representing the Rajan Lab this year. Shout out to co-first authors @satpreetsingh.bsky.social and @neurostrow.bsky.social.

Catch their sessions to learn about our newest work 🧠🤖

#NeuroAI #CompSci

25 5 0 0

(hollaback/cc)

@drun.v25media.com

3 days ago

Tony Hoare (1934-2026) Turing Award winner and former Oxford professor Tony Hoare passed away last Thursday at the age of 92. Hoare is famous for quicksort, ALGO...

Tony Hoare has left the building.

blog.computationalcomplexity.org/2026/03/tony...

#compsci #history #csp

2 1 0 0

merely rhys

@rhys.mastodon.rhys.wtf.ap.brid.gy

3 days ago

Tony Hoare (1934-2026) _Turing Award winner and former Oxford professorTony Hoare passed away last Thursday at the age of 92. Hoare is famous for quicksort, ALGOL, Hoare logic and so much more. Jim Miles gives his personal reflections._ --- Jill Hoare, Tony Hoare, Jim Miles. Cambridge, 7 September 2021 Last Thursday (5th March 2026), Tony Hoare passed away, at the age of 92. He made many important contributions to Computer Science, which go well beyond just the one for which most Maths/CompSci undergraduates might know his name: the quicksort algorithm. His achievements in the field are covered comprehensively across easy-to-find books and articles, and I am sure will be addressed in detail as obituaries are published over the coming weeks. I was invited in this entry to remember the Tony that I knew, so here I will be writing about his personality from the occasions that I met him. I visited Tony Hoare several times in the past 5 years, as we both live in Cambridge (UK) and it turned out that my family knew his. As a Mathematics graduate, I was very keen to meet and learn about his life from the great man himself. I was further prompted by a post on this blog which mentioned Tony a few times and summarised a relevant portion of his work. I took a print out of that entry the first time I visited him to help break the ice - it is the green sheet of paper in the picture above. Tony read the entry and smiled, clearly recalling very well the material of his that it referenced, and then elaborating a bit, explaining how vastly programs had scaled up in a rather short space of time and how they typically require different methods than many of those he had been developing in the early days. I was aware that Tony had studied Classics and Philosophy at university so I was keen to learn how one thing had led to another in the development of his career. He explained that after completing his degree he had been intensively trained in Russian on the Joint Services School for Linguists programme and was also personally very interested in statistics as well as the emerging and exciting world of computers. This meant that after his National Service (which was essentially the JSSL) he took on a job 'demonstrating' a type of early computer, in particular globally, and especially in the Soviet Union. He described the place of these demonstrations as 'fairs' but I suppose we might now call them 'expos'. In a sense, this seemed like a very modest description of his job, when in fact - reading up on Tony's career - he was also involved in the development of code for these devices, but perhaps that's a historical quirk of the period: being a demonstrator of these machines meant really knowing them inside and out to the point of acting on the dev team (AND, one might deduce, being fluent in Russian!). Tony would tell these stories with a clarity and warmth that made it clear that certainly he was still entirely 'all there' mentally, and that his memory was pinpoint sharp, even if there were some physical health issues, typical for anyone who makes it so far into their 80s (and, as we now know, beyond!). A story that I was determined to hear from the source was the legendary quicksort 'wager'. The story goes that Tony told his boss at Elliott Brothers Ltd that he knew a faster sorting algorithm than the one that he had just implemented for the company. He was told 'I bet you sixpence you don't!'. Lo and behold, quicksort WAS faster. I asked Tony to tell this story pretty much every time we met, because I enjoyed it so much and it always put a smile on both of our faces. To his credit, Tony never tired of telling me this story 'right from the top'. I had hoped to visit again in the past year and record him telling it so that there was a record, but unfortunately this did not happen. However, I discover that it is indeed recorded elsewhere. One detail I might be able to add is that I asked Tony if indeed the wager was paid out or if it had merely been a figure of speech. He confirmed that indeed he WAS paid the wager (!). A detail of this story that I find particularly reflective of Tony's humble personality is that he went ahead and implemented the slower algorithm he was asked to, while he believed quicksort to be faster, and before chiming in with this belief. It speaks to a professionalism that Tony always carried. About 50% of our meetings were spent talking about these matters relating to his career, while the rest varied across a vast range of topics. In particular, I wanted to ask him about a story that I had heard from a relative, that Tony - whilst working at Microsoft in Cambridge - would like to slip out some afternoons and watch films at the local Arts Picturehouse. This had come about because on one occasion a current film in question was brought up in conversation and it transpired Tony had seen it, much to the bemusement of some present. The jig was up - Tony admitted that, yes, sometimes he would nip out on an afternoon and visit the cinema. When I met Tony and gently questioned him on this anecdote he confirmed that indeed this was one of his pleasures and his position at Microsoft more than accommodated it. On the topic of films, I wanted to follow up with Tony a quote that I have seen online attributed to him about Hollywood portrayal of geniuses, often especially in relation to Good Will Hunting. A typical example is: "Hollywood's idea of genius is Good Will Hunting: someone who can solve any problem instantly. In reality, geniuses struggle with a single problem for years". Tony agreed with the idea that cinema often misrepresents how ability in abstract fields such as mathematics is learned over countless hours of thought, rather than - as the movies like to make out - imparted, unexplained, to people of 'genius'. However, he was unsure where exactly he had said this or how/why it had gotten onto the internet, and he agreed that online quotes on the subject, attributed to him, may well be erroneous. One final note I would like to share from these meetings with Tony is perhaps the most intriguing of what he said, but also the one he delivered with the greatest outright confidence. In a discussion about the developments of computers in the future - whether we are reaching limits of Moore's Law, whether Quantum Computers will be required to reinvigorate progress, and other rather shallow and obvious hardware talking points raised by me in an effort to spark Tony's interest - he said 'Well, of course, nothing we have even comes close to what the government has access to. They will always be years ahead of what you can imagine'. When pressed on this, in particular whether he believed such technology to be on the scale of solving the large prime factorisation that the world's cryptographic protocols are based on, he was cagey and shrugged enigmatically. One wonders what he had seen, or perhaps he was engaging in a bit of knowing trolling; Tony had a fantastic sense of humour and was certainly capable of leading me down the garden path with irony and satire before I realised a joke was being made. I will greatly miss this humour, patience, and sharpness of mind, as I miss everything else about Tony. RIP Tony Hoare (11 January 1934 - 5 March 2026)

CAR Hoare has passed. Damn.

#tech #Compsci

blog.computationalcomplexity.org/2026/03/tony-hoare-1934-...

0 0 0 0

@leighjohnston.bsky.social

4 days ago

I predict that human generated data will be the most sought after commodity in the post AI "singularity": artists currently worrying about AI have nothing to fear.
Human generated data will kick AI model collapse into the long grass.
#AI #compsci #IRC

0 0 0 0

@leighjohnston.bsky.social

4 days ago

we need to teach AI how to get drunk so they can leverage the Ballmer Peak. #AI #compsci

0 0 0 0

Jason Moore

@moorejh.bsky.social

4 days ago

'Shock! Shock! I learned yesterday that an open problem I’d been working on for several weeks had just
been solved by Claude Opus 4.6. It seems that I’ll have to revise my opinions about “generative AI” one of these days' ~ Don Knuth cs.stanford.edu/~knuth/paper... #claude #opus #ai #compsci

7 1 0 0

IDEATE

@ideatecommunity.bsky.social

6 days ago

IDEATE Webinar - Universal Design for Learning in tech Join Dr Anne Ozdowska for a webinar to understand Universal Design for Learning in computer science and software engineering courses at Australian universities.

🚨 LAST CHANCE TO REGISTER!

Our FREE webinar on Universal Design for Learning in computing educators at Australian universities starts tomorrow 2pm AEDT!

Secure your spot now:
events.humanitix.com...

#UDL #TechEducation #SoftwareEngineering #CompSci #HigherEd #Accessibility

0 0 0 0

@leighjohnston.bsky.social

1 week ago

interestingly the most valuable commodity in the future will be non-AI generated data, I predict. #AI #compsci #modelcollapse #IRC

0 0 0 0

@anthrog1rl.bsky.social

1 week ago

#Science #Chemistry #CompSci

0 0 0 0

IDEATE

@ideatecommunity.bsky.social

1 week ago

IDEATE Webinar - Universal Design for Learning in tech Join Dr Anne Ozdowska for a webinar to understand Universal Design for Learning in computer science and software engineering courses at Australian universities.

Standard coding curricula don’t always fit every student 👩‍💻👨‍💻

Level up your teaching with IDEATE's free webinar on Universal Design for Learning (UDL) for computing courses

🗓️ Tuesday, 10 March
🎟️events.humanitix.com...

#CompSci #EngineeringEducation #UDL #TeachingTech #InclusiveDesign

0 0 0 0

@leighjohnston.bsky.social

1 week ago

"C/C++" vs Rust

#cpp #rust #coding #compsci #gamedev

5 1 0 0

Craig’n’Dave

@craigndave.bsky.social

1 week ago

Endianness: the byte order battle in computing! Big-endian vs little-endian — why does it matter? Find out in our new blog! 🥚💻

Read here 👉 craigndave.org/what-is-endi...

#TeamCompSci #CraignDave #Coding #CompSci

0 0 0 0

Velma ($400/2000)

@verymuchvelma.bsky.social

1 week ago

Join the SegFault Union Discord Server! SegFault Union is a casual place for programmers to get together and learn and build together. We've also got casual off topic channels, so whether you write code or not, feel free to come hang out. |...

I probably haven't said this enough yet, but come join our discord!

I run a small #tech focused community on discord called "Segfault Union"

No matter what your skill level come hang out!

discord.gg/jas22Y4tBH

#SoftwareDeveloper
#SoftwareEngineer
#CompSci
#Programmer
#Developer

0 0 0 0

Jeff Horton :canada:

@jeffhorton.mstdn.ca.ap.brid.gy

1 week ago

Original post on mstdn.ca

In a video I'm watching

"This is not uncommon problem in more serious situations, but it's being framed as if we are helping a thief, but remember that's just fun algorithms is not about breaking the law, algorithms doesn't help burglary, but we are framing it as a funny instance, and sometimes […]

0 1 0 0

@leighjohnston.bsky.social

2 weeks ago

Have to say: ChatGPT is great for C++ code reviews. #AI #compsci #cpp #gamedev

2 1 0 0

Michael Martinez :verified:

@michael-martinez.c.im.ap.brid.gy

2 weeks ago

Original post on c.im

Computer science students looking to graduate this year: if you're not experimenting with vibe coding, you should be. This is an opportunity for you to learn about the limitations and flaws in AI-generated code (and there are TON of them). If you start interviewing for jobs with some experience […]

0 1 1 0

@leighjohnston.bsky.social

2 weeks ago

if you fire all the junior devs and use AI instead who is going to replace the senior devs who haven't drunk the AI koolaid? #irc #AI #compsci

0 0 0 0

Wes Fryer

@wesfryer.com

2 weeks ago

A slide titled "Scratch Maze Tournaments 24-25 February 2026" displaying tournament information. The top section features a two-column table with "Bracket link" options for "Computer Programming 2(6)" and "Computer Programming 1(1)". The bottom section is divided into three parts: an interactive spinning wheel on the left featuring the Scratch cat logo and various participant names with links, large blue hyperlinked text reading "Spinner Mazes" in the center, and a right-hand column listing multiple Scratch project URLs under the header "Maze Link".

Our Scratch Coding Maze Tournament starts TODAY!
docs.google.com/document/d/e...

Full lesson details on:
lessons.wesfryer.com/lessons/scra...

#coding #Scratch #create2learn #CompSci #ComputerScience #ComputationalThinking

1 0 0 0

@leighjohnston.bsky.social

3 weeks ago

AI is great but you have to be able to call on everything you learnt prior to the arrival of LLMs to leverage the stochastic parrots properly (i.e. fixing their errors and flat out hallucinations): this is why pure vibe coding by people with no coding experience remains bullsh*t. #compsci #AI

0 0 0 0

@leighjohnston.bsky.social

3 weeks ago

New reality for today's software developer/programmer/engineer: he/she/they only have to research what an AI has generated if he/she/they have to fix what AI has generated, which in my experience is not an uncommon situation. #coding #AI #compsci

0 0 0 0

GödelEscherSpock 🦿

@adam-hill.bsky.social

3 weeks ago

Very cool visualization of important tokens (and later on attention heads) for a LLMs incoming token stream.

#ai #compsci

3 0 0 0