L
@lingodotdev
Which AI can actually be trusted to write code unsupervised?
17
0
13
5.3K
Community Sentiment Analysis
Real-time analysis of public opinion and engagement
Sentiment Distribution
75% Engaged
41% Positive
34% Negative
Positive
41%
Negative
34%
Neutral
25%
Critical Perspectives
Community concerns and opposing viewpoints
1
Many replies push back hard on the claim that AI will replace developers, calling it hype and even “ridiculous,” with sarcasm and dismissive jokes
Many replies push back hard on the claim that AI will replace developers, calling it hype and even “ridiculous,” with sarcasm and dismissive jokes.
2
Several commenters praise AI as a productivity boost—examples like “80% of the code” being generated—but stress the final 20% requires human fixes for production readiness
Several commenters praise AI as a productivity boost—examples like “80% of the code” being generated—but stress the final 20% requires human fixes for production readiness.
3
Writers emphasize that AI can write code but can’t decide what’s worth building or fully understand problem context; humans remain the architects of purpose
Writers emphasize that AI can write code but can’t decide what’s worth building or fully understand problem context; humans remain the architects of purpose.
4
Security and quality worries recur
people point to vulnerabilities, inefficient generated code, and the need for careful human review before deployment.
5
Many predict new job shapes—prompt engineering, AI training, and reviewer roles—where humans shift from typing syntax to specifying and validating outcomes
Many predict new job shapes—prompt engineering, AI training, and reviewer roles—where humans shift from typing syntax to specifying and validating outcomes.
6
Views range from pragmatic optimism to flat rejection
some are hopeful about progress, others say it’s “not gonna happen” and that developers aren’t going anywhere for years.
7
Specific model experiences surface
Opus and Claude Code get credit for working with existing codebases but are criticized for failing to create robust projects from scratch.
8
The tone mixes skepticism, humor, and defensiveness—frequent one-liners, insults, and eye-rolls underscore a strong demand for human oversight
The tone mixes skepticism, humor, and defensiveness—frequent one-liners, insults, and eye-rolls underscore a strong demand for human oversight.
B
@bygregorr
wn me that tools like Claude Code are changing the game, but they still need a human touch to understand the problem, not just write the code. This shift will likely create new jobs in areas like prompt engineering and AI training. Humans will focus on what to build, not just h
6
0
0
900
O
@odinsbadeye
"Wow! So who is going to prompt the AI now?" "95% of the same people who were writing code before." "Oh"
3
0
0
418
Supporting Voices
Community members who agree with this perspective
1
Many replies insist that “humans writing syntax is over”, arguing the bottleneck has shifted to intent, architecture, and judgment. Commenters emphasize a move up the abstraction ladder
intent → constraints → systems → outcomes, not faster typing.
2
A strong thread reports this change is already happening inside teams — LLM-assisted commits, agentic workflows, and IDEs that coordinate subagents are described as commonplace
Users point to concrete experience (hundreds of AI-driven commits, tools like CaffeineAI) as evidence the transition accelerated quicker than expected.
3
There's palpable worry about job displacement and grief
predictions of significant layoffs (figures like “250k” were invoked), people feeling forced into management roles, and critiques of corporations prioritizing profit over employment. Some see poetic irony in engineers becoming the architects of the models that replace them.
4
Many replies treat Ryan Dahl’s remarks as especially credible — authority lends weight to the claim that coding-as-typing is finished
His name is invoked as a signal that this is not mere hype but a real paradigm shift.
5
Respondents sketch the emerging skill set
writing clear requirements, system design, editing AI output, and coordinating agents. There’s even talk of a new, machine-oriented programming language and IDEs optimized for orchestration rather than keystroke-level authoring.
6
A mix of optimism and caution appears
some celebrate new opportunities (entrepreneurial wins, faster progress), while others call for focus on safety, regulation, and critical domains before automating everything (health, voting systems, anti-fraud, etc.).
7
A few voices are defiant or bemused — from gloating about skipping degrees to quips that coding will be a history lesson — but most contributions converge on the same point
the practical nature of software work is changing, and skills will need to follow.
D
@davidpattersonx
It's interesting that the computer science people are putting themselves out of work first. The developers of AI have put a special focus on coding and math. Other professions will fall just as quickly as the focus shifts.
61
0
4
5.0K
T
@thecsguy
ryan dahl spent a decade trying to fix node with deno and then just decided to replace the developers instead. the ultimate refactor
29
0
0
3.2K
A
@ashebytes
there is a grief to accept w this https://t.co/m5Kg83n4qj
18
0
2
5.0K