Ali

Time to face the truth, nothing is more permanent than a temporary solution that works.

[Ali @ginandconic]

@apostylee Google takeout diye bir hizmeti var Google’ın oraya girdikten sonra iCloud seçip hesap bilgilerinizi girdikten sonra aktarıyor hepsini 60bin foto iki gün civari sürüyor

[Ali @ginandconic]

@ozgurcd Çok çok iyi, Portekizli dostlarımızı bile alıştırdık.

[Ali @ginandconic]

@ottomanenglish Those 3 items are the only cheaper things in the north, everything else is almost 1.5 - 2x the price in south.

[Ali @ginandconic]

@GokcePehlvnoglu Tarzindan da hic odun vermiyor. 30 yildir ayni cicekli gomlek, hastasiyiz...

[Ali @ginandconic]

@DoanDadelen5 vaultwarden, bugsink

[Ali @ginandconic]

@fkadev Elgato facecam

[Ali @ginandconic]

@fkadev https://t.co/8060Vzvb9K

[Ali @ginandconic]

@fkadev IKEA BÄSTBOLL

[Ali @ginandconic]

@kartolsu Biz Aegean ile uçtuk 8.1 kilo geldi su tosuncuk ve çantası

[Ali @ginandconic]

@karthikponna19 Ada

[Ali @ginandconic]

@treshpink İki öğün yemek parası

[Ali @ginandconic]

@populicc Bizde böyle

[Ali @ginandconic]

@mdisec Türkiye cloudflare stream servisini zaten engellemiş

[Ali @ginandconic]

@azmimengu Drops ve memrise şahane uygulamalar

[Ali @ginandconic]

1993: basic
2005: C#
2006: C#
2007: Actionscript
2008: Java
2009: Java
2010: Java
2011: Java
2012: Java
2013: Java
2014: Php
2015: Php
2016: Php
2017: Php
2018: Php
2019: Php
2020: JavaScript
2021: JavaScript
2022: typescript
2023: typescript
2024: typescript
2025: typescript

[boris @boristane]

my main programming language:

2008 c
2009 c++
2010 c++
2011 matlab
2012 c++
2013 c++
2014 matlab
2015 matlab
2016 vba
2017 python
2018 javascript
2019 typescript
2020 typescript
2021 typescript
2022 typescript
2023 typescript

note: I was in school til 2015

[Ali @ginandconic]

@gecersandin

[Ali @ginandconic]

@rellivdev @ersincmt Aynen oyle, yedekler de ayni binadaymis

[Ali @ginandconic]

@Nerraograyghost Kara borsa biletler can yakıyor ama kesinlikle değer!

[Ali @ginandconic]

@imTolstoyevski yeni bölüm gelir mi ?

[Connor Davis @connordavis_ai]

This MIT paper just broke my brain.

Everyone keeps saying LLMs can't do real logical reasoning. Turns out we've just been teaching them wrong this whole time.

These researchers built something called PDDL-INSTRUCT that actually teaches models to think through planning problems step by step. Not just pattern matching - actual logical reasoning.

Here's how it works:

Phase 1: show the model correct and incorrect plans with explanations. Basic stuff. Phase 2 is where it gets interesting. They make the model generate explicit reasoning for every single action, then use an external verifier to check if each step is logically sound.

The numbers are wild. Llama-3-8B jumped from 28% to 94% accuracy on planning benchmarks. That's not incremental improvement - that's a completely different capability emerging.

What's smart is they don't trust the model to check its own work. They use VAL, a formal planning verifier, to validate every logical step. When the model screws up, it gets specific feedback about exactly what went wrong.

The two-stage training is clever. First stage focuses purely on better reasoning chains. Second stage optimizes for actually solving the problem. This prevents the model from just gaming the metrics.

One finding caught my attention - detailed feedback destroys binary feedback. Just telling a model "wrong" vs explaining exactly which preconditions failed makes a huge difference. The gap is especially big on complex problems.

This isn't trying to replace symbolic planners. It's teaching neural networks to reason like symbolic planners while keeping external verification. That's actually sustainable.

The implications go way beyond planning. Any multi-step reasoning task could benefit from this approach. We might finally be seeing how to teach LLMs structured thinking instead of just sophisticated autocomplete.

Makes me wonder what other "impossible" capabilities are just sitting there waiting for the right training approach.