Eggb4by

332 posts

Eggb4by

Eggb4by

@eggbaby

Entrepreneur. Indie Developer. Creative Thinker. https://t.co/wKGuIH37ev,AI Token Calculator https://t.co/a0GFbH5YYc, For my daughter

Katılım Şubat 2009
380 Takip Edilen79 Takipçiler
Ethan Mollick
Ethan Mollick@emollick·
Thats good for the dumb reason that I am burning tokens having Sol in Codex playing Slay the Spire 2’s daily challenge (so randomized rules, today it is Ascension 3 Defect with hoarder) to see if it can It just beat the Act 1 boss. Quite surprised. I’ll check back in the morning
Tibo@thsottiaux

Introducing... another usage limit reset for all our ChatGPT Work and Codex users. Should land over next 30 minutes. Hope you have an awesome weekend. Thank you for pushing our systems to the absolute limit, we have never seen traffic increase so quickly. Keep the feedback coming and we'll keep shipping.

English
23
3
344
65.1K
Eggb4by
Eggb4by@eggbaby·
@AranYogesh Coding agents are quietly becoming infrastructure.
English
0
0
0
1
Eggb4by
Eggb4by@eggbaby·
@LLMJunky Tiny feature. Huge quality-of-life upgrade.
English
0
0
0
1
am.will
am.will@LLMJunky·
Okay this is really cool. Codex app has a new feature that is really useful. Highlight any text, and you'll notice a new popup. "More Details" uses GPT 5.5 Instant to explain in more detail about the highlighted portion of your text, allowing you to understand the reasoning/concept/statement/etc in greater detail. What's interesting about this versus a side chat is that it's instantaneous, and still retains context from the thread. With /side chat, you'd have to type out a prompt "what are the details on this?" With "more details" the prompt is implicit and the response is instantenous Very nice feature, as I'm always making side chats/branches to expand/learn/understand agent outputs better.
am.will tweet media
English
30
24
328
32.4K
Eggb4by
Eggb4by@eggbaby·
@pvncher New model, old prompting habits. Every time.
English
0
0
0
1
eric provencher
eric provencher@pvncher·
With 5.6 Sol, a lot of people are still prompting the model exactly as they did 5.5 It's important to note that 5.6 Sol is a lot more tenacious and thorough than previously models. Check out the guide I wrote here to get better outcomes learn.chatgpt.com/docs/prompting
English
61
199
2.2K
197.4K
Eggb4by
Eggb4by@eggbaby·
@CFGeek The crash isn't the scary part. The recovery plan is.
English
0
0
0
1
Eggb4by
Eggb4by@eggbaby·
@sama Two resets in 24 hours? Okay, today we benchmark in production.
English
0
0
0
1
Eggb4by
Eggb4by@eggbaby·
@rasbt The annoying part is pricing pages still present this as a model choice. Effort level can change the answer more than the model name.
English
0
0
0
1
Sebastian Raschka
Sebastian Raschka@rasbt·
For agentic coding, one can say: - Unless you need Terra Ultra perf, it's always better to use a Luna model with higher effort setting (same or better performance but cheaper). - Forget everything below Sol High, use Luna with higher effort settings here - Forget Sol Extra High, use Terra Ultra here - The extra cost of Sol Ultra is probably not worth it over Max
Sebastian Raschka tweet media
English
318
423
4.4K
818.6K
Eggb4by
Eggb4by@eggbaby·
@gdb Being able to interrupt and ask "why does this matter?" is the useful part. Otherwise it's just another audio feed.
English
0
0
0
1
Greg Brockman
Greg Brockman@gdb·
GPT-Live for personalized news briefings:
Simon Smith@_simonsmith

I discovered a great use case for GPT-Live today: Personalized, interactive news briefings. I have a scheduled task in ChatGPT to create a personalized newscast. I’ve been having Chat read that aloud using the text-to-speech button on replies. But today I wondered: Can GPT-Live read this to me and then I can ask it questions as it’s going, and then it resumes where it left off? The answer is yes! You can just hit the GPT-Live button in an existing thread and talk about it, and that includes a scheduled task thread. You can ask questions, ask for additional research, and do anything else GPT-Live allows, and it will address your request and then resume reading where it left off. This is really cool.

English
36
17
418
87.3K
Eggb4by
Eggb4by@eggbaby·
@emollick The voice got natural faster than the reasoning did. You notice the gap as soon as the conversation gets hard.
English
0
0
0
2
Ethan Mollick
Ethan Mollick@emollick·
The new ChatGPT voice is quite impressive to use, really worth a minute to try it out on your phone. (while staying aware that the voice model is not going to be as smart as a full thinking model)
English
42
26
425
128.9K
Unsloth AI
Unsloth AI@UnslothAI·
We’re releasing new Qwen3.6 quants that run 2.5× faster on your GPU. Qwen3.6-27B NVFP4 runs on 24GB VRAM. 35B-A3B can hit 17,561 tok/s (B200). We also improved accuracy, tool calling, agent use, and looping. Guide: #nvfp4" target="_blank" rel="nofollow noopener">unsloth.ai/docs/models/qw… Qwen3.6 NVFP4: huggingface.co/collections/un…
Unsloth AI tweet media
English
167
344
2.7K
321.2K
Eggb4by
Eggb4by@eggbaby·
@dair_ai This is the part model-only evals miss. Retries and bloated context can erase a model gain before the user sees it. I'd like to see harness configuration reported alongside every agent benchmark.
English
0
0
0
8
DAIR.AI
DAIR.AI@dair_ai·
// The Harness Effect // (bookmark it) Now more that ever pay very close attention to the orchestration harness and its effect on costs and performance. This study ran 22 evaluation tasks on six foundation models (Claude Sonnet 4.6, Gemini 3.1, Qwen 3.6, GLM 5.1, and others), then change only the orchestration layer. Holding models constant, the harness cuts blended cost per task 41%, tokens per task 38%, and median wall-clock 44%, with completion quality at parity. Two results do the work. Efficiency is model-invariant, every model gets 33 to 61% cheaper. Quality gain correlates almost perfectly with baseline model strength (r=0.99 across six models), a effect they call harness leverage. Why does it matter? On this workload the orchestration layer moved cost per task more than the full spread of the model menu did. The harness is the one component whose efficiency multiplies across every model an organization runs. Paper: arxiv.org/abs/2607.06906 Learn to build effective AI agents in our academy: academy.dair.ai
DAIR.AI tweet media
English
17
34
237
18.3K
Eggb4by
Eggb4by@eggbaby·
@omarsar0 The 1.3B version is what got me. Curious what "coherent for an hour" means in practice: visual consistency, persistent world state, or both?
English
0
0
0
8
elvis
elvis@omarsar0·
Cool open-source release. LingBot-World 2.0 holds 720p at 60 fps in real time and stays coherent for a full hour of interaction. A 1.3B variant runs on a single consumer GPU, so you can actually run a world model at home. Fully open source, weights and harness on Hugging Face. Worth a look.
English
7
4
20
11.6K
Eggb4by
Eggb4by@eggbaby·
@colifran_ The part I'd want to test is forgetting. When a source changes, does OpenWiki know what to update or remove? Proactive memory gets messy fast if old context keeps winning.
English
1
0
0
29
Colin Francis
Colin Francis@colifran_·
agent memory has always been reactive. OpenWiki makes it proactive. connect to sources, tell it what you care about, and your agent hits the ground running . what we're building is really exciting! try it out and get involved, it's open source! 👇 github.com/langchain-ai/o…
English
1
5
34
22.1K
Eggb4by
Eggb4by@eggbaby·
@emollick For a lot of knowledge work, the audit trail is part of the deliverable. Hiding sources and intermediate decisions behind a "simpler" UI removes the control people actually need.
English
0
0
0
3
Ethan Mollick
Ethan Mollick@emollick·
Hint for all AI Labs as they branch out from work for programming to general knowledge work: non-coders are not just dumber coders Taking away a bunch of options from your coding app does not make it better for knowledge work. We need more types of control & visibility, not less
English
44
36
727
48.1K
Eggb4by
Eggb4by@eggbaby·
@arcprize 7.8% sounds small until you look at what changed. Sol can read the scene now, then still gets stuck once the inference chain gets deeper. That failure mode tells me more than another clean benchmark win.
English
0
0
0
7
ARC Prize
ARC Prize@arcprize·
GPT-5.6 Sol sets a new SOTA on ARC-AGI-3: 7.8% Sol is the first verified frontier model to ever beat an ARC-AGI-3 game It is the best model at orienting in a situation it's never encountered
ARC Prize tweet media
English
91
292
2.8K
517.8K
Geek
Geek@geekbb·
- Claude Fable 5 - OpenAI GPT-5.6 - Grok 4.5 - Meta Muse Spark 1.1 - Hy3 应接不暇……
English
5
0
7
7.1K
Eggb4by
Eggb4by@eggbaby·
现在电脑里有一个ChatGPT和一个ChatGPT Classic....图标还是一样的，自己都晕了
中文
0
0
0
27
Eggb4by
Eggb4by@eggbaby·
Chatgpt的Google插件也更新了
Eggb4by tweet media
中文
0
0
0
16

Keşfet

@emollick @AranYogesh @LLMJunky @pvncher @CFGeek @sama @rasbt @gdb