This is the funniest #AI news I heard all week. I almost wish I was on the @microsoft.com @discord.com just to witness it. www.ibtimes.com/microsoft-di.... #AIHype Shout out to @i-dont-like-ai.bsky.social for calling it to my attention. Long live #microslop!
Do you just say professional vibe coding?
Everyone talks about "vibe coding" which means an unskilled person generating code. What is it called when a skilled person uses the #AI as their team of junior developers, giving instructions on what to build and reviewing code in a really knowledgeable and thoughtful manner?
Maybe Sam Altman needs a better haircut 😆
Apparently #AI is pretty unpopular. (Except for programmers who generally love it.) Why do you think that is?
This is in comparison to the long history of previous tech booms including mobile phones, the internet, eletricity, the automobile, etc.
Gift article link:
www.nytimes.com/2026/02/21/t...
Are you feeling the #AI burn-OUT? Everyone is coding so fast with Claude-Code and Cursor that you are SWAMPED IN PR REVIEWS? You are not alone!
The Harvard Business Review reports AI intensifies work:
hbr.org/2026/02/ai-d...
TechCrunch reports on the burnout:
techcrunch.com/2026/02/09/t...
On the bright side, that means the rest of the economy was actually doing better than everyone thought last year, right? The article doesn't really explain that.
DM me if you can't get the full text.
The analysts who came up with this failed to account for the fact that 75% of new data center spending is on chips and foreign manufactured equipment - that means it was other countries' growth, not the US. Seems like a kind of hard fact to miss, you know what I mean?
The story that #AI produced so much growth in 2025 because of data center spending? Turns out it was too good to be true.
#AIHype
www.washingtonpost.com/technology/2...
Are LLM's like brains? Scientists at UCLA tackled the question and wrote this paper: Illusions of Alignment Between Large Language Models and Brains Emerge From Fragile Methods and Overlooked Confounds TLDR: No similarity between LLM's and real brains. #AIHype #LLM
www.biorxiv.org/content/10.1...
Robustness issues cut across domains but are particularly well-studied in benchmark-based evaluations and social reasoning, where minor, semantically-preserving perturbations can lead to large and inconsistent shifts in model outputs.
Application-specific limitations cluster in certain domains: for example, Theory of Mind instability in implicit social reasoning, and inability to generalize to novel Math Word Problem structures in formal reasoning.
Fundamental failures – stemming from intrinsic architectural or training constraints – manifest across all reasoning types. For example, cognitive biases such as confirmation bias.
Diagram of LLM reasoning failures. Robustness issues, application-specific failures, and fundamental failures are further categorized into subcategories.
LLM reasoning still has so many flaws that Stanford and Caltech researchers wrote an entire journal article categorizing them! 🤯
The paper is in Transactions on Machine learning Research, available on Arxiv: arxiv.org/pdf/2602.06176
#LLM #LLMHype #AI #AIHype #AIBubble
By 2030 will AI lead to:
“Unemployment increased significantly.” : No
“A breakthrough treatment or cure for a major disease.”: Yes
“A major global security event.”: Yes
How likely is it that we will see A.G.I. in the next 10 years?: Unlikely
In the near term, What will A.I.’s impact be on
medicine? : Moderate
programming?: Large
scientific research?: Moderate
transportation?: Small
education?: Moderate / Large
mental health?: Moderate
art and creativity?: Moderate / Large
The The New York Times asked 8 "leading thinkers" on the direction AI is taking us in the near and long term. TLDR? Here is a quick summary of the consensus...
Gift link for the full article: www.nytimes.com/interactive/...
The new yorker article: www.newyorker.com/culture/2025... (DM me if you want to see the full text) For a nice summarization and expansion on the arguments, check out this blog post by Srinivas Rao skooloflife.medium.com/the-agentic-...
Why? LLM based Agents just know how to talk, and talking and doing are miles apart. The devil is in the details of execution. LLM's don't actually "understand" anything - they just predict the next likely text. That leaves too many opportunities for a task to go off the rails and never recover.
This is acknowledges that the initial hype and optimism was overblown. This is exactly as predicted by myself and people smarter than me, such as Gary Marcus. A New Yorker writer on the east coast see this, but so many people I speak to here in silicon valley are still drunk on the agent kool aid?
The New Yorker had an great article by Cal Newport mocking the silicon valley fascination with "Agents" in 2025. AI industry leaders started 2025 saying "2025 is the year of the agent" and by years end had backtracked into "2025 is the start of the decade of the agent" #LLM #Agentic #AIHhype #AI
Maybe #GenAI really is just memorizing, says new research - it seems to be a form of "lossy compression". This actually may have a big impact on the copyright debate and legal challenges.
www.theatlantic.com/technology/2...
Also noteworthy: credit default swaps on Oracle jumped when they borrowed to build data centers.
In my past life I was a fixed income (bond) analyst, so this very interesting to me, at the intersection of #AI and #Fixedincome. 🤓 #aihype #aibubble
Bond investors are more risk averse than stock investors so they call out problems in business models. To build AI data centers investors are requiring rates above 9% coupon on bonds. - almost double the interest for similar non-data center bonds. #aibubble #aihype
www.nytimes.com/2025/12/26/b...
The good news? The speed of code generation meant overall output is up. And AI generated code does have fewer typos. 🙂
Rather than displacing human programmers, human programmers shift into AI management and reviewing, which keeps them pretty busy.
Are you worried about the quality of #AI generated code? A new study showed that AI generated code had 1.7X more issues, including 1.4X more critical issues.
www.techradar.com/pro/security...
#LLM #codegeneration #AIHype
Happy New Year. Save half site-wide at manning.com. Fireworks and xmas ornaments in the background
Happy New Year! If you're still online reading this, you can still save 50% in @manning.com's EOY sale. 2026 will be an important year to Fight #Churn With #DataScience and #MachineLearning. www.manning.com/books/fighti...
Beats me. I remember going to Vancouver twice. I think there were some Montreal years? Then the more recent incarnation in San Diego. I actually went to one of the San Diego NeurIPS in the 20-teens. I did go to an SFN in New Orleans. Maybe you were thinking of that?
I went back before it was cool too! I went in Vancouver in the early aughts. For better or worse, I’m still making my career in this field, which has blown up so much.
DM me for a full article if you can't get it from the link.
www.wsj.com/tech/ai/ai-c...
