Lazizbek Ergashev | IT Blog
Open in Telegram
568
Subscribers
No data24 hours
+27 days
+230 days
Loading data
Similar Channels
Tags Cloud
No data
Any problems? Please refresh the page or contact our support manager.
Incoming and Outgoing Mentions
---
---
---
---
---
---
Attracting Subscribers
October '26Oct '26
October '26
+3
in 0 channels
September '260
in 0 channels
Get PRO
August '260
in 0 channels
Get PRO
July '260
in 0 channels
Get PRO
June '260
in 0 channels
Get PRO
May '260
in 0 channels
Get PRO
April '260
in 0 channels
Get PRO
March '260
in 1 channels
Get PRO
February '260
in 5 channels
Get PRO
January '260
in 0 channels
Get PRO
December '250
in 0 channels
Get PRO
November '250
in 0 channels
Get PRO
October '250
in 0 channels
Get PRO
September '250
in 0 channels
Get PRO
August '250
in 1 channels
Get PRO
July '250
in 0 channels
Get PRO
June '250
in 0 channels
Get PRO
May '250
in 0 channels
Get PRO
April '250
in 1 channels
Get PRO
March '250
in 5 channels
Get PRO
February '250
in 0 channels
Get PRO
January '250
in 0 channels
Get PRO
December '24
+11
in 0 channels
Get PRO
November '24
+16
in 0 channels
Get PRO
October '24
+17
in 0 channels
Get PRO
September '24
+11
in 0 channels
Get PRO
August '24
+4
in 0 channels
Get PRO
July '240
in 0 channels
Get PRO
June '240
in 2 channels
Get PRO
May '240
in 2 channels
Get PRO
April '240
in 0 channels
Get PRO
March '24
+4
in 0 channels
Get PRO
February '24
+21
in 1 channels
Get PRO
January '24
+25
in 1 channels
Get PRO
December '23
+15
in 2 channels
Get PRO
November '23
+22
in 0 channels
Get PRO
October '23
+78
in 5 channels
Get PRO
September '23
+31
in 0 channels
Get PRO
August '23
+27
in 0 channels
Get PRO
July '23
+27
in 0 channels
Get PRO
June '23
+14
in 0 channels
Get PRO
May '23
+37
in 0 channels
Get PRO
April '23
+26
in 0 channels
Get PRO
March '23
+30
in 0 channels
Get PRO
February '23
+86
in 0 channels
Get PRO
January '23
+117
in 0 channels
Get PRO
December '22
+41
in 0 channels
Get PRO
November '22
+46
in 0 channels
Get PRO
October '22
+95
in 0 channels
Get PRO
September '22
+31
in 0 channels
Get PRO
August '22
+127
in 0 channels
Get PRO
July '22
+55
in 0 channels
Get PRO
June '22
+51
in 0 channels
Get PRO
May '22
+65
in 0 channels
Get PRO
April '22
+36
in 0 channels
Get PRO
March '22
+68
in 0 channels
Get PRO
February '22
+33
in 0 channels
Get PRO
January '22
+33
in 0 channels
Get PRO
December '21
+38
in 0 channels
Get PRO
November '21
+33
in 0 channels
Get PRO
October '21
+51
in 0 channels
Get PRO
September '21
+312
in 0 channels
Get PRO
August '21
+46
in 0 channels
Get PRO
July '21
+67
in 0 channels
Get PRO
June '21
+121
in 0 channels
Get PRO
May '21
+52
in 0 channels
Get PRO
April '21
+427
in 0 channels
| Date | Subscriber Growth | Mentions | Channels | |
| 18 October | 0 | |||
| 17 October | 0 | |||
| 16 October | 0 | |||
| 15 October | +1 | |||
| 14 October | +1 | |||
| 13 October | +1 | |||
| 12 October | 0 | |||
| 11 October | 0 | |||
| 10 October | 0 |
Channel Posts
Dasturlashni 2016-yilda, 14 yoshimda boshladim.
16 yoshimda frilanser bo'lib ishlay boshladim, chunki u yoshdagi bolani hech kim ishga olmasdi, o'shanda O'zbekistonda dasturlash hozirgidek mashhur ham emasdi. Shuning uchun yoshimni 20-22 deb yozib, xorijiy frilans platformalarda ish topdim. Ishlarim shunchalik yaxshi ketdiki, freelancehunt.com saytida O'zbekistondan #1 dasturchi bo'ldim.
Hozir frilansni to'xtatdim, chunki u yerni AI botlar bosib ketdi. 9-5 ishlaydigan oddiy ishga o'tdim. Nolimayman: ish barqaror, kelajakni rejalashtirsa bo'ladi.
Bilaman, ko'pchiligingiz kanalga boshida yozgan frilans maslahatlarim uchun qo'shilgansiz. Lekin hozir bunaqa maslahatlar yozishda ma'no ko'rmayapman. Hech kim maslahat so'ramaydi, o'zim ham bu yo'nalishda rivojlanishni to'xtatdim, chunki endi siz ham, men ham AIdan so'raymiz.
Bu AI degan narsa qayerga olib borishini bilmayman, lekin maslahatim: moslashuvchan bo'ling va o'zgarishlarga har qachongidan ham ochiq bo'ling. AIdan ko'p foydalanaman, lekin unga shubha bilan qarayman. U insoniyatni yaxshiroq kelajakka olib boradi deb o'ylamayman. U faqat korporatsiyalarning foydasini oshiradi, xolos deb o'ylayman.
Part [2/2]
🧠 @lazerg| 2 | Bugun 25 yoshga to'ldim. Chorak asr degani bu, shuning uchun shu paytgacha nimalarga ulgurganimni yozib qo'ymoqchiman.
22 yoshda uylandim, hozir qizimiz bor. Kichik oilamiz bilan deyarli 10 ta davlatni ko'rdik. 🇮🇩 Indoneziyada oddiy baliqchilarning uyida yashashdan boshlab to dunyodagi eng rivojlangan davlat 🇸🇬 Singapurni ham ko'rdik. O'tgan yili bolalikdagi orzuimni amalga oshirib, digital nomad bo'ldim. Lekin o'ylaganimdek qiziq emas ekan, shuning uchun 4-5 oydan keyin O'zbekistonga qaytib keldik.
Ha, plyajda bir qo'limda noutbuk, bir qo'limda kokos bilan ishladim. Lekin bu internetda ko'rsatilgandek emas, ko'pi feyk. Aslida juda noqulay. Tashqarida noutbuk qizib ketadi, batareya tez o'tiradi. Quyoshda esa yorqinlikni oxirigacha qo'ysangiz ham ekranda hech narsa ko'rinmaydi. Va hokazo.
To'rt tilda gaplashaman: 🇹🇯 tojik, 🇺🇿 o'zbek, 🇷🇺 rus va 🇬🇧 ingliz. Bu juda complex: Buxoroda tojikcha gaplashamiz, maktabda, litseyda o'zbekcha o'qiganman, ishda esa rus va ingliz tilida ishlayman. Xullas, universalman ~
TATUda o'qidim. O'sha kunlar hayotimdagi eng yomon kunlardan edi, lekin muhim qadam ham bo'ldi. Hozir qandayligini bilmayman, lekin men o'qigan paytda tizim juda kringe edi: o'qishingdan ko'ra kiyimingga ko'proq qarashardi. Dekanning gapi hali ham esimda: "Qanchalik aqlli bo'lishingning farqi yo'q. Kostyum bryuk kiymasang, O'zbekiston uchun bir tiyin odamsan." 🤡
Ishim sharofati bilan Prezidentimiz Shavkat Mirziyoyev, vazir Sherzod Shermatov va boshqa ajoyib insonlar bilan uchrashganman.
Part [1/2]
🧠 @lazerg | 220 |
| 3 | Kanal muallifini 24 yoshida yo‘qotdik. 😔😔
Barcha ma’lumotlar izohlarda. | 290 |
| 4 | Korporatsiyalar AI'ni ishlatib bizni yo'qotmoqchi edi, qasos uchun biz AI'ni ishlatib o'zlarini yo'qotamiz.
😏
🧠 @lazerg | 315 |
| 5 | Adobe’ning barcha dasturlarini vibecode qilib tashlab, open source qilib qo‘yishibdi:
https://getartcraft.com/apps
🥀 adobe 💔
🧠 @lazerg | 629 |
| 6 | Upwork haqida
🧠 @lazerg | 524 |
| 7 | @khamroev gapim xato ekanini izohlarda isbotlab qo‘ydilar. Kalta o'ylab yozganim uchun uzr... 😭
🧠 @lazerg | 782 |
| 8 | 💡 O‘z millating va tarixing bilan maqtanish xuddi chorshanba kuni tug‘ilganing bilan maqtanishga o‘xshaydi: ikkalasiga ham tasodifan erishgansan.
#stopMillatchilik
🧠 @lazerg | 757 |
| 9 | ℹ️ Do not judge people by their clothes when you know nothing about their choices.
🧠 @lazerg | 660 |
| 10 | Biror narsada tanilib qolsangiz, vaqt o'tgani sayin o'sha sohada yanada kuchayib borasiz.
Dasturchi "X bilan shug'ullanadigan odam" sifatida tanilib qolsa, unga X bo'yicha loyihalar taklif qilina boshlaydi, X mavzusidagi konferensiya va meetuplarga chaqiriladi, X ga qiziquvchi yosh dasturchilar bilan tanishtiriladi, X bo'yicha kod review so'rashadi va X bo'yicha hamkorlik takliflari kela boshlaydi.
Reputatsiya imkoniyatlarni aynan sizga yo'naltiradi. Shu tariqa, asta-sekin siz X sohasida haqiqatan ham nodir darajada chuqur tajriba to'playsiz.
Tajriba reputatsiya yaratadi. Ammo reputatsiya ham ba'zida tajriba yaratadi.
🧠 @lazerg | 587 |
| 11 |  Solving a Rubik's Cube with graph theory.
🧠 @lazerg | 493 |
| 12 | Btw, I joined dmed.uz as Backend Engineer...
🧠 @lazerg | 672 |
| 13 |  Xuddi shunday, aka. AI open source narsalardan foydalanib, ishni qoyillatib qo‘yadiku, to‘g‘rimi?
Masalan, Node.js misolida tushuntiray, chunki buni ko‘proq odam biladi. AI’ga “backend qilib ber” desangiz, nima qiladi? Open source narsalarni ishlatib qiladi. Masalan, backendni NestJS’da yozadi (u ham open source), database sifatida PostgreSQL’dan foydalanadi (u ham open source), keyin Prisma ORM’ni qo‘shadi (u ham open source) va oxirida response qaytaradi.
Keyin hammasini Docker’ga joylaydi (u ham open source), Ubuntu serverga deploy qiladi (u ham open source) va bo‘ldi. “Ana, AI backend qildi”, deb o‘ylaysiz. Aslida esa AI shunchaki open source loyihalarning documentation’larini o‘qib, ularni bir-biriga ulab berdi.
Lekin biroz low-level’ga tushsak-chi? PostgreSQL ichkarida qanday ishlaydi? Nega request ko‘p vaqt olyapti? Men, masalan, oyiga 10 million odam ishlatadigan app’ning backendini shunchaki vibe coding qilib qurib bo‘lishiga umuman tasavvur qilolmayman.
Chunki PostgreSQL’da qaysi column uchun qanday data structure’li index ishlatish kerakligini bilmasangiz, 10 mln user uchun tez resposne qaytaradigan backend tuzolmaysiz. Yoki cachinglarni chuqur bilmasangiz. Prosta "cache the backend" emas, qanday qilsa optimizatsiyalandi..
Vibecoderlar va chala dasturchilar odatda shitty startupcha narsalar qilishadi, uni o‘zidan va do‘stlaridan boshqa hech kim ishlatmaydi. Keyin esa: “Ana, AI qildi, demak dasturlashni bilish shart emas”, deyishadi.
Lekin scalability-chi? Dasturingiz 1 million userga chiqqanda ham ishlaydimi? (agar chiqsa albatta) | 932 |
| 14 | Menga eng alam qilayotgani o‘zim hurmat qilgan dasturchilarni ko‘rsam, avvallari ular juda qiziqarli videolar va dasturlash bo‘yicha foydali darslar qilishardi. Hozir esa qarasam, deyarli hammasi faqat AI haqida gapiryapti. Qaysi model kuchli, Astra zo‘rmi, Fable zo‘rmi va hokazo. Yana bittasi chiqib: “Mana, endi kompyuterga o‘tirishim ham shart emas, telefondan turib kod yozaman”, deb maqtanadi.
Dasturlashni chuqur o‘rgatadigan odamlar qolmayotgandek. Hamma AI bo‘yicha nimanidir o‘rgatyapti. Kodni coding agentlar yozib berayotgani bilan dasturlashni chuqur tushunish endi kerak emas bo‘lib qolmadiku. Aksincha bu hozirgi AI world da juda muhim.
Bunga oddiy misol sifatida Владилен Минин (https://www.youtube.com/@VladilenMinin) ni keltirishim mumkin. JavaScript va Vue.js’ni aynan shu kishidan o‘rganganman. U kishining kurslari shunchaki “.map nima?”, “.filter nima?” degan darajada emasdi. JavaScript ichkarida qanday ishlashi, Promise’lar nega asinxron ishlayotgandek ko‘rinishi, aslida qanday ishlashi va buni qanday isbotlash mumkinligi kabi narsalarni chuqur tushuntirardi.
Hozir esa kanalini ko‘rsam, asosan “AI agentim buni qiladi”, “AI agentim uni qiladi” degan mavzular. U kishini ham tushunsa bo‘ladi, albatta. Bozordagi talabga qarab moslashishadi. Dasturlash haqida chuqur kurs chiqarishsa, ehtimol prosmotr kamroq bo‘lardi.
Lekin baribir dasturlashni chuqur o‘rgatadigan kontent juda kamayib ketayotgandek tuyuladi. Ehh, bilmadim...
🧠 @lazerg | 593 |
| 15 | Kuzatganimcha, vibe coderlar va chala dasturchilar o‘ynayotgan shitty startupcha-startupcha o‘yinidan foyda qilayotganlar faqat AI modellar (OpenAI, Anthropic…) va infrastructure (domen, server, Vercel, Supabase…) sotayotganlar.
🧠 @lazerg | 467 |
| 16 | Being always available when your boss writes you is not normal.
Even an API has rate limits.
🧠 @lazerg | 509 |
| 17 | https://youtu.be/o3NXZdTnYYk
🧠 @lazerg | 582 |
| 18 | Biz MustaqiLLM’ni yaratdik!
Bugun biz uchun muhim kun, o‘zbek tili uchun noldan yaratilgan Katta Til Modelini taqdim etdik.
Bu mavjud xorijiy modelning oddiy fine-tune qilingan versiyasi emas.
Biz modelni noldan yaratdik va o‘qitdik.
Model:
→ 5 milliard parametr
→ 40 milliard tokenlik ko‘p tilli dataset
→ 48 ming tokenli BPE tokenizer
→ O‘zbek tiliga maxsus moslashtirilgan
→ Grouped-Query Attention
→ QK-Normalization
→ Uzun kontekst bilan ishlashga tayyor arxitektura
Nega mustaqil LLM muhim?
AI bugungi raqamli infratuzilmaning muhim qismiga aylanmoqda. Shuning uchun o‘z tilimiz va ma’lumotlarimiz bilan ishlaydigan, texnologik jihatdan o‘zimiz nazorat qila oladigan modellarni yaratish strategik ahamiyatga ega.
Bu nafaqat o‘zbek tilining AI’dagi rivoji, balki lokal AI ekotizimi, ma’lumotlar xavfsizligi va kelajakdagi AI texnologiyalariga mustaqil hissa qo‘shish uchun ham muhim.
MustaqilLLM hali yakuniy manzil emas. Uning bilim, fikrlashi va umumiy til qobiliyatlarini yanada rivojlantirishda davom etamiz.
Ammo bu o‘zbek tilidagi mustaqil AI’ni yaratish yo‘lidagi muhim boshlanish.
© NeuronAI Jamoasi
Model vaznlari huggingface platformasida joylashtirildi: https://huggingface.co/NeuronUz/MustaqiLLM
@adizbek_ergashev | 708 |
| 19 | 🪔 Elektr shunaqa darajada oddiy narsa bo‘lib ketdiki, faqat puli yo‘qlar sham ishlatadi.
AI shunaqa darajada oddiy narsa bo‘lib ketadiki, faqat puli yo‘qlar miyasini ishlatadi.
🧠 @lazerg | 809 |
| 20 |  Even if I don't want to disclose the details, the layoff probably affected me too... What I can say is: It was not my decision
🧠 @lazerg | 1 017 |
