29 Jan 2025, 16:56 UTC≈9,870 views25 reactionsread 6 August 2026
Где все клоуны кто задвигал про dpo) Мне буквально СТО стартапа юникорна полтора года назад задвигал что RL больше не нужен, так как есть DPO😂
🔥17😘5💘2🥰1
Channel
AGI
@higgsfield
On this record: Growth · Engagement · Reactions · Posts · Citations · Cite this entry
433subscribers
+7 since we began measuring on 6 August 2026
Risers and fallers across the register · movement among entries of Under 1,000.
Register entry
| Telegram ID | -1001822271916 |
|---|---|
| Type | Channel |
| Username | @higgsfield |
| Created | 17 November 2022 — measured — dated from the channel’s first post |
| First recorded | 6 August 2026 |
| Last confirmed live | 14 August 2026 |
| Measurements held | 3 |
| Confirmed unchanged | 1 time, most recently 14 August 2026 |
| On Telegram | t.me/higgsfield |
Growth
| Measured (UTC) | Subscribers | Change |
|---|---|---|
| 14 Aug 2026, 22:15 | 433 | +7 |
| 6 Aug 2026, 19:30 | 426 | no change |
| 6 Aug 2026, 06:05 | 426 | first reading |
Engagement
15 posts held, back to 17 November 2022 — the reader has not yet reached the start of this channel’s public history, so older posts may sit further back, unread. Read across 1 pageof Telegram’s post history, 20 posts per page.
Nothing published in the last 30 days. ERR and ER are rolling 30-day measures, so there is nothing to compute — we hold 15 posts for this entry, the most recent from 29 January 2025. An engagement rate over an empty window would be a number about nothing.
Reaction mix
109 reactions across 12 posts, in 11 distinct kinds. The most used accounts for 60.6% of them.
| Reaction | Count | Share | Share, drawn |
|---|---|---|---|
| 🔥 | 66 | 60.6% | |
| ❤ | 14 | 12.8% | |
| 😘 | 9 | 8.26% | |
| 🥰 | 7 | 6.42% | |
| 👍 | 4 | 3.67% | |
| 💘 | 3 | 2.75% | |
| 👏 | 2 | 1.83% | |
| ❤🔥 | 1 | 0.917% | |
| 💋 | 1 | 0.917% | |
| 💯 | 1 | 0.917% | |
| 😍 | 1 | 0.917% |
No sentiment is inferred, and none should be read in. This table is ordered by count and by nothing else. Emoji do not carry stable meaning across languages or communities — 🙏 is thanks in one channel and mourning in another — so we publish which ones were pressed and how often, and pass no judgement on what an audience meant by them.
Precision. Telegram publishes reaction counts per emoji and short-forms each one — 4.34K, 1.2M — so any single kind at or above 1,000 reaches us at three significant figures, and only counts below 1,000 are exact. The shares above are ratios of those figures and carry the same error. This is also why the total here can differ slightly from a reaction total printed elsewhere on the page: both are sums of the same rounded parts, taken over samples with different edges.
Coverage. Reactions were read on 12 of the 15 sampled posts in this sample. Summed by Telegram’s own count on each post — not by adding up the per-emoji breakdown above — those same posts carry 109reactions in total: the kind of figure the paragraph above means by “a reaction total printed elsewhere on the page”.
Measured over the 15 most recent posts we hold, published 17 November 2022 to 29 January 2025, using the newest reading held for each. Telegram Stars are excluded: they are a payment, not a reaction, and they have their own section.
Recent posts
17 May 2023, 20:18 UTC≈10,500 views10 reactionsread 6 August 2026
Numbers every LLM Developer should know by Ray and Jeff Dean https://github.com/ray-project/llm-numbers
❤5👍3🔥1😘1
16 May 2023, 21:23 UTC≈8,570 views9 reactionsread 6 August 2026
Еще один template language для LLM, теперь от Microsoft. Потихонько девелопмент LLM applications стабилизируется. Люди понимают что решения типо langchain не юзабельны для сложных композиции LLM, так как у вас получается огромный лапша код полностью не юзабельный, особенно если он пишется большой командой. Идея использовать классические template language с LLM достаточно очевидна, поэтому в скором времени мы увидим …
🔥5❤2😘2
16 May 2023, 09:49 UTCviews —
Channel photo updated
16 May 2023, 09:00 UTC≈5,730 views11 reactionsread 6 August 2026
На данный момент молодому ресерчеру(или deeptech стартапу ахах) важно сфокусироваться на 4 направлениях 1. Engineering - Gpu optimisation/distributed training - методы по типу fast attention которые позволили обучить модели типо gpt-4 с большим context length 2. Lora и другие методы адаптеров и файн тюна, в будущем скорее всего у каждого человека будет персональный АИ который будет tailored к нему, вопрос где держа…
🔥7🥰2❤1👍1
16 May 2023, 08:34 UTC≈4,010 views10 reactionsread 6 August 2026
В первом посте я писал на сколько важно валидировать аутпут LLM. Это позволяет убрать галлюцинирование, контролировать constraints, и самое главное строить сложные композиции из LLM агентов которые смогут общаться между собой, декомпозировать задачи, критиковать и рефайнить. Например у нас есть LLM аналитик, LLM инженер, LLM board member итд После разговора с одним VC, он говорит что за последнее время видел 50 LLM …
🔥6🥰2❤1💯1
14 May 2023, 21:41 UTC≈3,350 views7 reactionsread 6 August 2026
Photo
Компании в будущем будут полностью управляться AI с минимальным вмешательством человека. Интересный пейпер про операционную аналитику, ERP, digital twins и LLM Towards autonomous system: flexible modular production system enhanced with large language model agents https://arxiv.org/pdf/2304.14721.pdf
🔥5❤1🥰1
14 May 2023, 21:37 UTC≈2,960 views2 reactionsread 6 August 2026
Комбинирование constraint SAT solver-a c LLM (in context learning) Reliable Natural Language Understanding with Large Language Models and Answer Set Programming https://arxiv.org/pdf/2302.03780.pdf
❤🔥1🔥1
14 May 2023, 20:25 UTC≈2,840 views4 reactionsread 6 August 2026
Хороший survey по prompt engineering написал Lil’Log из openai. Где он уложил 25 ключевых пейперов по prompt engineering каждый в пару предложении https://lilianweng.github.io/posts/2023-03-15-prompt-engineering/
🔥3😘1
14 May 2023, 20:23 UTC≈2,360 views6 reactionsread 6 August 2026
Все пейперы выше по prompt engineering, это нельзя назвать разделом machine learning, так как это другая дисциплина. Порог входа в понимание пейперов очень низкий, вам не нужно знать хорошо математику или computer science. Поэтому можно читать сотни таких пейперов без напряга
🔥5🥰1
14 May 2023, 20:21 UTC≈2,340 views7 reactionsread 6 August 2026
Интересные пейперы по prompt engineering которые прочитал на этих выходных. From Words to Code: Harnessing Data for Program Synthesis from Natural Language https://arxiv.org/pdf/2305.01598.pdf Unstructured and structured data: Can we have the best of both worlds with large language models? https://arxiv.org/pdf/2304.13010.pdf TaskMatrix.AI: Completing Tasks by Connecting Foundation Models with Millions of APIs htt…
🔥5❤1💘1
14 May 2023, 20:13 UTC≈1,720 views5 reactionsread 6 August 2026
Photo
Как формальные методы могут помочь прунить спейс если объяснить на пальцах? Например мы хотим генерить код, в данном случае мы можем на каждом этапе генерации токена проверять удовлетворяет ли корректному синтаксису полученная строка. Другой пример если мы ставим какой то constraint на аутпут LLM. Например a && b, понятно если a=false, нет смысла дальше проверять эту ветку дерева. Математически мы используем произв…
🔥4💋1
Showing the 12 most recent of 15 posts we hold for @higgsfield. View and reaction counts are the latest single reading for each post, not a live figure, and a recent post is still accumulating both. A view count marked ≈ was rounded by Telegram before we ever saw it — t.me prints views in full below 1,000 and to three significant figures above, so ≈1,200,000 means somewhere between 1,150,000 and 1,249,999. Unmarked counts are exact. Text is reproduced from the public post preview and truncated for length.
Citation-graph rank
Citation-graph rank — 145,180 of 1,481,243entries in the measured graph. A weighted position computed from the forward and mention edges below — republished posts weigh more than named mentions — and recomputed periodically, over the whole graph. Published only as this ordinal position, never as a score: a position is a fact, and a score printed beside one channel’s name would read as a verdict this register does not make. The two counts beneath stay separate for the same reason mentions are never summed with forwards anywhere else on this page — a named-by count costs nothing to manufacture. The top 100 by this measure, or how it is computed.
Mentions
Named by 7 registered channels — every channel on the register whose own posts have named this one, by its current username or any other username it currently holds, merged from two separately captured readings of the same fact so a namer caught by only one of them is not missed and a namer both caught is not counted twice. A username this channel has since dropped is not matched — that handle may belong to someone else now, and crediting today’s namer to yesterday’s owner would misattribute it.
Named by
Channels on the register whose posts name this channel's handle.
Vocabulary English
@LearnVocabularyEnglish_2022 · 7,8344 postsPhoto_Tvoritel_Hub
@photo_tvoritel_hub · 31,7514 postsPredmet AI
@predmet_ai · 3,8934 postsNARUTO
@naruto_nf · 5,8182 postsarmandá
@armanda_a · 1,5091 postДенис про Контент-завод | Автоматизация | Вайбкодинг
@autogenerator_reels · 7,3561 postVBB SOUND
@vbbsound · 421 post
@LearnVocabularyEnglish_2022 · 7,8344 postsPhoto_Tvoritel_Hub
@photo_tvoritel_hub · 31,7514 postsPredmet AI
@predmet_ai · 3,8934 postsNARUTO
@naruto_nf · 5,8182 postsarmandá
@armanda_a · 1,5091 postДенис про Контент-завод | Автоматизация | Вайбкодинг
@autogenerator_reels · 7,3561 postVBB SOUND
@vbbsound · 421 post
A mention is a weaker signal than a forward and is counted separately for that reason — naming a channel is not republishing it, and a handle in a post body is easy to place deliberately. The post counts beside each row below are distinct posts in which the handle appeared, from posts we have read on both sides — the “Named by N registered channels” figure above is a different count, of distinct NAMING CHANNELS rather than posts, and is not the sum of the rows under it.
Cite this entry
A live page changes as we take new readings, so a citation should name the measurement it is based on, not just the URL. The line below cites the subscriber count as measured 14 August 2026 — this entry's latest reading, not the date you are reading this.
“AGI” (@higgsfield), 433 subscribers as measured 14 August 2026. Telegram Register, tgregister.com/channel/higgsfield.
Full measurement history, CC BY 4.0. Every reading this register holds for this entry, not just the latest one, as a dated, downloadable record: CSV · JSON. Free to use with attribution to tgregister.com. Each file carries its own generation timestamp, which is the figure to cite for exactly when the data was retrieved.