fsndz

6 ms·

31.
by fsndz 2y ago
SAP is kinda of a monopoly no ? The software is bloated and complex to use. The UI is dated but everyone still uses it because it is so embedded into the core arteries of businesses.
32.
by fsndz 2y ago
this is top notch commentary
33.
by fsndz 2y ago
So why are people so excited about MCP, and so suddenly? I think you know the answer by now: hype. Mostly hype, with a bit of the classic fascination among software engineers for architecture. You just say Model Context Protocol, server, cl
34.
by fsndz 2y ago
Nice release. I see that everyone is playing the differentiation game now: https://medium.com/thoughts-on-machine-learning/llama-4-and-...
35.
by fsndz 2y ago
vibe coding is overrated. The good sequence is: learn to code, then vibe code when possible. Vibe coding is a trap for junior software engineers. https://www.lycee.ai/blog/why-vibe-coding-is-overrated
36.
by fsndz 2y ago
the problem of agents is not the lack of standards, but reliability (reliability of tool use and reliability of outcomes). MCP does not solve any of that.
37.
by fsndz 2y ago
so we no longer need langchain and stuff like that, that's a win. But MCP also feels a bit overrated: https://www.lycee.ai/blog/why-mcp-is-mostly-bullshit
38.
by fsndz 2y ago
indeed. open source will win. sam Altman was wrong: https://www.lycee.ai/blog/why-sam-altman-is-wrong
39.
by fsndz 2y ago
it's just a marketing stunt like prompt engineer with 200K salary back in the day. And the fact that you are talking about it and sharing the link to the company's profile means the stunt is working.
40.
by fsndz 2y ago
show me an example of such a complex thing obtained after just writing prompts
41.
by fsndz 2y ago
vibe coding has been in the scene since the emergence of ChatGPT perhaps...
42.
by fsndz 2y ago
it's definitely nice for prototypes, but nothing more complex.
43.
by fsndz 2y ago
the worst programmer nowadays is the vibe coder. vibe coding is so overrated imo https://www.lycee.ai/blog/why-vibe-coding-is-overrated
44.
by fsndz 2y ago
some people say agents can help. but still difficult to pull it without robust human in the loop: https://medium.com/thoughts-on-machine-learning/build-a-text...
45.
by fsndz 2y ago
I actually tried sth like that with the Bible back in 2021. scaling is bitch. very difficult to train these types of models.
46.
by fsndz 2y ago
this is an awesome article, I really enjoyed reading it.
47.
by fsndz 2y ago
DeepSeek Really Shattered OpenAI to Its Core. It’s funny because a few weeks back, in response to Elon Musk’s attempt to buy OpenAI, Sam Altman responded: ‘I wish he would just compete by building a better product.’ https://mediu
48.
by fsndz 2y ago
which hoops ?
49.
by fsndz 2y ago
DeepSeek really shook them to their core. Now they go for regulatory capture. Such a huge disappointment. Open source AI will win: https://medium.com/thoughts-on-machine-learning/the-laymans-...
50.
by fsndz 2y ago
it's really awesome !
51.
by fsndz 2y ago
yeah, sure, I guess the investors selling NVIDIA's stock like crazy know nothing about jevons
52.
by fsndz 2y ago
super impressive. we won't need that many GPUs in the future if we can have the performance of DeepSeek R1 with even less parameters. NVIDIA is in trouble. We are moving towards a world of very cheap compute: https://medium
53.
by fsndz 2y ago
just check moravec paradox: https://en.wikipedia.org/wiki/Moravec's_paradox
54.
by fsndz 2y ago
the thing is LLMs don't even do the kind of generalisations the dumbest human can do. while simultaneously doing some stuff the smartest human probably can't
55.
by fsndz 2y ago
Exactly. That's basically the problem with a lot of the current paradigm, they don't allow true generalisation. That's why some people say there won't be any AGI anytime soon: https://www.lycee.ai/blog&#x
56.
by fsndz 2y ago
Intelligence is not the bottleneck to progress or effectiveness in most companies. That’s why having access to more intelligence — human or artificial — will not solve everything.
57.
by fsndz 2y ago
yes the problem is that there is no consensus about what AGI should be: https://medium.com/@fsndzomga/there-will-be-no-agi-d9be9af44...
58.
by fsndz 2y ago
most humans are generally intelligent but can't do what you just said AGI should do...
59.
by fsndz 2y ago
it's so over, pretraining is ngmi. maybe sam Altman was wrong after all ? https://www.lycee.ai/blog/why-sam-altman-is-wrong
60.
by fsndz 2y ago
if Claude Code offers a better experience, users will rapidly move from cursor to Claude Code. Claude is for Code: https://medium.com/thoughts-on-machine-learning/claude-is-fo...

More ›