Ebenezer Don 🐘

Continued Pretraining: How specialiazed AI models are made

This is a good time to curate your TL and be intentional about who you follow and the contents that comes to you directly. I've done the work for you, just follow all these people. [ arranged in no particular order ] Engineering : @marlene_zw @unicodeveloper @czarify @sarah_edo @cassidoo @akinkunmi @ASpittel @AdamRackis @chiziaruhoma @Samson_Goddy @bolaabanjo @saniyusuf @matteocollina @coder_blvck @honour_can_code @cassiecodes @AdoraNwodo @SQLBrethen @chinedu_10 @ireaderinokun @tomisin_dev @Abiodun0x @mykeels @iChuloo @SRunsewe @chidiwilliams__ @Timmysofine @kvlly @Sisinerd @bleso_a @Geektutor - You'll get alot of community info. @julian__duru @Didicodes @edesiri @musti_ag @silasadedoyin @lanreadelowo @ebenezerDN

@hackSultan @marlene_zw @unicodeveloper @czarify Solid list 🙏🔥🔥

GPT-5.6 Is Incredible. I Hate Using It youtu.be/lF6xgolXN68

@claudeai @AnthonyLimo Lol

We're extending Claude Fable 5 access on all paid plans, as well as keeping Claude Code’s weekly rate limits 50% higher, through July 19.

Seeing trillion dollar companies fumble very basic functionalities. Most of these platforms were much better before the LLM boom. AI has made us too lazy.

Fable 5 is a good model. But unusable. Literally. At least we can use GPT 5.6. Anthropic lost this round. Cursor would have less users leaving their platform if they borrowed some of OpenAI's pricing model and "generosity" Still got the best agent in the game.

The leap from GPT 5.5 to 5.6 is the same as Opus 4.6 to 4.7 Literally a 0.1 leap Don't believe the hype

GPT 5.6 Sol is great at planning, finding issues/edge-cases and brainstorming. But it's really bad at working. It can go on and on for hours doing nothing and just burning tokens. You'll feel like it's being thorough, but it's not. You're just excited that it's wasting your time. Even for the simplest tasks, this model wastes time excruciatingly. The hype will wear off in a couple of days, and you'll realise that this model version is 5.6 and not 6 for a reason. And I say this as someone on the $200 pro plan. I've worked with it's highest and lowest forms of intelligence, consistently for the past 30 hours. It is a good model, but very much overhyped. Best way to work with it - have it do the planning, then assign the work to GPT 5.5. Thank me next week.

@Danibholie 🔥🔥🔥

JavaScript for VIPs by @ebenezerDN is still my go-to resource whenever I need a JavaScript refresher. The concepts are explained in a simple, easy-to-follow way, with well-thought-out examples that make them stick. Highly recommend it to anyone learning or revisiting JavaScript

@DalgleishGX @phonefuturist What are you on about? This is glassmorphism you apple fanboys

@phonefuturist His video is an extremely bad example but it is Liquid Glass

Suddenly everything with blur effect is liquid glass

@thsottiaux @OpenAIDevs Second message in the chat and getting this

@rodydavis @antigravity Adopt auto-review mode from Codex

What would you like to see improved, added or removed from @antigravity 👀 We are working hard on building the best product for developers and happy to close any critical gaps.

@scaling01 Cursor is xAI

Cursor just announced a new 1.5+ trillion parameter model pre-trained on over 100k GPUs. According to Cursor CEO it is: "as big as Opus and GPT" Yes, the cat is out of the bag and I can spoil the party. Opus 4.5 to 4.8 and GPT-5 to GPT-5.5 are not that big! (they are all below 2T params) This means that the current performance of GPT-5.5 and Opus 4.8 is achievable for open-source, because guess what else is that size? -> DeepSeek-V4-Pro The only moat is scaling. So far, Anthropic is the only lab that successfully made the jump to the ~10T scale, which is also why I no longer see OpenAI catching up until year end. Anthropic can just keep throwing RL compute at Mythos for the next 1-2 years and it will keep improving. Google went not quite as big and sparsity-maxxed a bit too much. They also have apparently no clue how to RL to make the model actually usable. OpenAI is still scarred from GPT-4.5 xAI and Meta are still planning.

@kskrygan Really wish JetBrains was at the forefront of this new era. But there's still time, and I hope to see you win!

Now it’s official. JetBrains remains the last big independent player in developer tooling. Our job is to deliver the most cost-efficient, deeply integrated, and genuinely enjoyable AI experience across our IDEs and beyond. We’re on it. Stay tuned.

@cursor_ai @SpaceX Congrats! 🎉

We're excited to join forces with @SpaceX to advance the frontier of useful AI. Expect significant improvements to Cursor soon.

BREAKING: The US government just took down Claude Fable 5 and it's worse than it looks Full video: youtu.be/Ll0c6xZttdQ